Search Results

Search found 4580 results on 184 pages for 'faster'.

Page 165/184 | < Previous Page | 161 162 163 164 165 166 167 168 169 170 171 172  | Next Page >

  • What Counts For a DBA: Fitness

    - by Louis Davidson
    If you know me, you can probably guess that physical exercise is not really my thing. There was a time in my past when it a larger part of my life, but even then never in the same sort of passionate way as a number of our SQL friends.  For me, I find that mental exercise satisfies what I believe to be the same inner need that drives people to run farther than I like to drive on most Saturday mornings, and it is certainly just as addictive. Mental fitness shares many common traits with physical fitness, especially the need to attain it through repetitive training. I only wish that mental training burned off a bacon cheeseburger in the same manner as does jogging around a dewy park on Saturday morning. In physical training, there are at least two goals, the first of which is to be physically able to do a task. The second is to train the brain to perform the task without thinking too hard about it. No matter how long it has been since you last rode a bike, you will be almost certainly be able to hop on and start riding without thinking about the process of pedaling or balancing. If you’ve never ridden a bike, you could be a physics professor /Olympic athlete and still crash the first few times you try, even though you are as strong as an ox and your knowledge of the physics of bicycle riding makes the concept child’s play. For programming tasks, the process is very similar. As a DBA, you will come to know intuitively how to backup, optimize, and secure database systems. As a data programmer, you will work to instinctively use the clauses of Transact-SQL DML so that, when you need to group data three ways (and not four), you will know to use the GROUP BY clause with GROUPING SETS without resorting to a search engine.  You have the skill. Making it naturally then requires repetition and experience is the primary requirement, not just simply learning about a topic. The hardest part of being really good at something is this difference between knowledge and skill. I have recently taken several informative training classes with Kimball University on data warehousing and ETL. Now I have a lot more knowledge about designing data warehouses than before. I have also done a good bit of data warehouse designing of late and have started to improve to some level of proficiency with the theory. Yet, for all of this head knowledge, it is still a struggle to take what I have learned and apply it to the designs I am working on.  Data warehousing is still a task that is not yet deeply ingrained in my brain muscle memory. On the other hand, relational database design is something that no matter how much or how little I may get to do it, I am comfortable doing it. I have done it as a profession now for well over a decade, I teach classes on it, and I also have done (and continue to do) a lot of mental training beyond the work day. Sometimes the training is just basic education, some reading blogs and attending sessions at PASS events.  My best training comes from spending time working on other people’s design issues in forums (though not nearly as much as I would like to lately). Working through other people’s problems is a great way to exercise your brain on problems with which you’re not immediately familiar. The final bit of exercise I find useful for cultivating mental fitness for a data professional is also probably the nerdiest thing that I will ever suggest you do.  Akin to running in place, the idea is to work through designs in your head. I have designed more than one database system that would revolutionize grocery store operations, sales at my local Target store, the ordering process at Amazon, and ways to improve Disney World operations to get me through a line faster (some of which they are starting to implement without any of my help.) Never are the designs truly fleshed out, but enough to work through structures and processes.  On “paper”, I have designed database systems to catalog things as trivial as my Lego creations, rental car companies and my audio and video collections. Once I get the database designed mentally, sometimes I will create the database, add some data (often using Red-Gate’s Data Generator), and write a few queries to see if a concept was realistic, but I will rarely fully flesh out the database since I have no desire to do any user interface programming anymore.  The mental training allows me to keep in practice for when the time comes to do the work I love the most for real…even if I have been spending most of my work time lately building data warehouses.  If you are really strong of mind and body, perhaps you can mix a mental run with a physical run; though don’t run off of a cliff while contemplating how you might design a database to catalog the trees on a mountain…that would be contradictory to the purpose of both types of exercise.

    Read the article

  • Private Cloud: Putting some method behind the madness

    - by Sudip Datta
    Finally, I decided to join the blogging community. And what could be a better time to start than the week after OpenWorld 2012. 50K+ attendees, demonstrations, speaker sessions and a whole lot of buzz on Oracle Cloud..It was raining clouds in this year's Openworld. I am not here to write about Oracle's cloud strategy in general, but on Enterprise Manager's cloud management capabilities. This year's Openworld was the first after we announced the 12c Cloud Control and we were happy to share the stage with quite a few early adopters. Stay tuned for videos from our customers and partners, I will post them as they get published. I met a number of platform administrators in Oracle-DBAs, Middleware Admins, SOA Admins...The cloud has affected them all, at least to the point where it beckoned more than just curiosity..Most IT infrastructure are already heavily virtualized (on VMWare and on others including Oracle VM), and some would claim they are already on “cloud” (at least their Sysadmins told them so). But none of them were confident of the benefits because their pain points continued to grow.. Isn't cloud supposed to ease those? Instead, they were chasing hundreds of databases running on hundreds of VMs, often with as much certainty propounded by Heisenberg. What happened to the age-old IT discipline around administration, compliance, configuration management? VMs are great for what they are. I personally think they have opened the doors to new approaches in which an application stack gets provisioned and updated. In fact, Enterprise Manager 12c is possibly the only tool out there that can provision full-fledged application as VM Assemblies. In this year's Openworld, customers talked on how they provisioned RAC and Siebel assemblies, which as the techies out there know, are not trivial (hearing provisioning time for Siebel down from weeks to hours was gratifying indeed). However, I do have an issue with a "one-size fits all" approach to cloud. In a week's span, I met several personas: Project owners requiring an EC2 like VM instance for their projects Admins needing the same for Sparc-Solaris. DBAs requiring dedicated databases for new projects APEX Developers needing just a ready-to-consume schema as a service Java Developers looking for a runtime platform QA engineers needing a fast clone of their production environment If you drill down further, you will end up peeling more layers of the details. For example, the requirements for Load testing and Functional testing are very different. For Load testing the test environment should ideally be the same as the production. You shouldn't run production on Exadata and load test on a VM; they will just not be good representations of one another. For Functional testing it does not possibly matter. DBAs seem to be at the worst affected of the lot. It seems they have been asked to choose between agile provisioning and  faster runtime performance. And in some cases, it is really a Hobson's choice, because their infrastructure provider made no distinction between the OLTP application and the Virtual desktop! Sad indeed. When one looks at the portfolio of services that we already offer (vanilla IaaS, VM Assembly based PaaS, DBaaS) or have announced (Java PaaS, Instant Cloning, Schema-aaS), one can possibly think that we are trying to be the "renaissance man" ! Well I would have possibly digested that had it not been for the various personas that I described above. Getting the use cases right is very important for an application such as cloud management. We iterate and iterate over these over and over again and re-validate them in CABs (Customer Advisory Boards). We consider over the major aspects of tenancy: service placement, resource isolation (can a tenant execute an expensive SQL and run away with all the resources), quota and security. We, in Engineering, keep reminding ourselves that we are dealing with enterprise clouds. We owe it to our customer base ! In the coming posts, I will drill down more into each of the services. In the meanwhile, here are some collateral and  demos for starters with EM 12c. http://www.oracle.com/technetwork/oem/cloud-mgmt/index.html Sudip Datta The views expressed here are my own and do not necessarily reflect the views of Oracle. Stay Connected: Twitter | Facebook | YouTube | Linkedin | Newsletter --

    Read the article

  • Developing Schema Compare for Oracle (Part 6): 9i Query Performance

    - by Simon Cooper
    All throughout the EAP and beta versions of Schema Compare for Oracle, our main request was support for Oracle 9i. After releasing version 1.0 with support for 10g and 11g, our next step was then to get version 1.1 of SCfO out with support for 9i. However, there were some significant problems that we had to overcome first. This post will concentrate on query execution time. When we first tested SCfO on a 9i server, after accounting for various changes to the data dictionary, we found that database registration was taking a long time. And I mean a looooooong time. The same database that on 10g or 11g would take a couple of minutes to register would be taking upwards of 30 mins on 9i. Obviously, this is not ideal, so a poke around the query execution plans was required. As an example, let's take the table population query - the one that reads ALL_TABLES and joins it with a few other dictionary views to get us back our list of tables. On 10g, this query takes 5.6 seconds. On 9i, it takes 89.47 seconds. The difference in execution plan is even more dramatic - here's the (edited) execution plan on 10g: -------------------------------------------------------------------------------| Id | Operation | Name | Bytes | Cost |-------------------------------------------------------------------------------| 0 | SELECT STATEMENT | | 108K| 939 || 1 | SORT ORDER BY | | 108K| 939 || 2 | NESTED LOOPS OUTER | | 108K| 938 ||* 3 | HASH JOIN RIGHT OUTER | | 103K| 762 || 4 | VIEW | ALL_EXTERNAL_LOCATIONS | 2058 | 3 ||* 20 | HASH JOIN RIGHT OUTER | | 73472 | 759 || 21 | VIEW | ALL_EXTERNAL_TABLES | 2097 | 3 ||* 34 | HASH JOIN RIGHT OUTER | | 39920 | 755 || 35 | VIEW | ALL_MVIEWS | 51 | 7 || 58 | NESTED LOOPS OUTER | | 39104 | 748 || 59 | VIEW | ALL_TABLES | 6704 | 668 || 89 | VIEW PUSHED PREDICATE | ALL_TAB_COMMENTS | 2025 | 5 || 106 | VIEW | ALL_PART_TABLES | 277 | 11 |------------------------------------------------------------------------------- And the same query on 9i: -------------------------------------------------------------------------------| Id | Operation | Name | Bytes | Cost |-------------------------------------------------------------------------------| 0 | SELECT STATEMENT | | 16P| 55G|| 1 | SORT ORDER BY | | 16P| 55G|| 2 | NESTED LOOPS OUTER | | 16P| 862M|| 3 | NESTED LOOPS OUTER | | 5251G| 992K|| 4 | NESTED LOOPS OUTER | | 4243M| 2578 || 5 | NESTED LOOPS OUTER | | 2669K| 1440 ||* 6 | HASH JOIN OUTER | | 398K| 302 || 7 | VIEW | ALL_TABLES | 342K| 276 || 29 | VIEW | ALL_MVIEWS | 51 | 20 ||* 50 | VIEW PUSHED PREDICATE | ALL_TAB_COMMENTS | 2043 | ||* 66 | VIEW PUSHED PREDICATE | ALL_EXTERNAL_TABLES | 1777K| ||* 80 | VIEW PUSHED PREDICATE | ALL_EXTERNAL_LOCATIONS | 1744K| ||* 96 | VIEW | ALL_PART_TABLES | 852K| |------------------------------------------------------------------------------- Have a look at the cost column. 10g's overall query cost is 939, and 9i is 55,000,000,000 (or more precisely, 55,496,472,769). It's also having to process far more data. What on earth could be causing this huge difference in query cost? After trawling through the '10g New Features' documentation, we found item 1.9.2.21. Before 10g, Oracle advised that you do not collect statistics on data dictionary objects. From 10g, it advised that you do collect statistics on the data dictionary; for our queries, Oracle therefore knows what sort of data is in the dictionary tables, and so can generate an efficient execution plan. On 9i, no statistics are present on the system tables, so Oracle has to use the Rule Based Optimizer, which turns most LEFT JOINs into nested loops. If we force 9i to use hash joins, like 10g, we get a much better plan: -------------------------------------------------------------------------------| Id | Operation | Name | Bytes | Cost |-------------------------------------------------------------------------------| 0 | SELECT STATEMENT | | 7587K| 3704 || 1 | SORT ORDER BY | | 7587K| 3704 ||* 2 | HASH JOIN OUTER | | 7587K| 822 ||* 3 | HASH JOIN OUTER | | 5262K| 616 ||* 4 | HASH JOIN OUTER | | 2980K| 465 ||* 5 | HASH JOIN OUTER | | 710K| 432 ||* 6 | HASH JOIN OUTER | | 398K| 302 || 7 | VIEW | ALL_TABLES | 342K| 276 || 29 | VIEW | ALL_MVIEWS | 51 | 20 || 50 | VIEW | ALL_PART_TABLES | 852K| 104 || 78 | VIEW | ALL_TAB_COMMENTS | 2043 | 14 || 93 | VIEW | ALL_EXTERNAL_LOCATIONS | 1744K| 31 || 106 | VIEW | ALL_EXTERNAL_TABLES | 1777K| 28 |------------------------------------------------------------------------------- That's much more like it. This drops the execution time down to 24 seconds. Not as good as 10g, but still an improvement. There are still several problems with this, however. 10g introduced a new join method - a right outer hash join (used in the first execution plan). The 9i query optimizer doesn't have this option available, so forcing a hash join means it has to hash the ALL_TABLES table, and furthermore re-hash it for every hash join in the execution plan; this could be thousands and thousands of rows. And although forcing hash joins somewhat alleviates this problem on our test systems, there's no guarantee that this will improve the execution time on customers' systems; it may even increase the time it takes (say, if all their tables are partitioned, or they've got a lot of materialized views). Ideally, we would want a solution that provides a speedup whatever the input. To try and get some ideas, we asked some oracle performance specialists to see if they had any ideas or tips. Their recommendation was to add a hidden hook into the product that allowed users to specify their own query hints, or even rewrite the queries entirely. However, we would prefer not to take that approach; as well as a lot of new infrastructure & a rewrite of the population code, it would have meant that any users of 9i would have to spend some time optimizing it to get it working on their system before they could use the product. Another approach was needed. All our population queries have a very specific pattern - a base table provides most of the information we need (ALL_TABLES for tables, or ALL_TAB_COLS for columns) and we do a left join to extra subsidiary tables that fill in gaps (for instance, ALL_PART_TABLES for partition information). All the left joins use the same set of columns to join on (typically the object owner & name), so we could re-use the hash information for each join, rather than re-hashing the same columns for every join. To allow us to do this, along with various other performance improvements that could be done for the specific query pattern we were using, we read all the tables individually and do a hash join on the client. Fortunately, this 'pure' algorithmic problem is the kind that can be very well optimized for expected real-world situations; as well as storing row data we're not using in the hash key on disk, we use very specific memory-efficient data structures to store all the information we need. This allows us to achieve a database population time that is as fast as on 10g, and even (in some situations) slightly faster, and a memory overhead of roughly 150 bytes per row of data in the result set (for schemas with 10,000 tables in that means an extra 1.4MB memory being used during population). Next: fun with the 9i dictionary views.

    Read the article

  • Sass interface in HTML6 for upload files.

    - by Anirudha
    Originally posted on: http://geekswithblogs.net/anirugu/archive/2013/11/04/sass-interface-in-html6-for-upload-files.aspx[This post is about experiment & imagination] From Windows XP (ever last OS I tried) I have seen a feature that is about send file to pen drive and make shortcut on Desktop. In XP, Win7 (Win8 have this too, not removed) just select the file right click > send to and you can send this file to many places. In my menu it’s show me Skype because I have installed it. this skype confirm that we can add our own app here to make it more easy for user to send file in our app. Nowadays Many people use Cloud or online site to store the file. In case of html5 drag and drop you need to have site opened and have opened that page which is about file upload. You need to select all  and drag and drop. after drag and drop file is simply uploaded to server and site show you on list (if no error happen). but this file upload is seriously not worthy since I have to open the site when I do this operation.   Through this post I want to describe a feature that can make this thing better.  This API is simply called SASS FILE UPLOAD API Through This API when you surf the site and come into file upload page then the page will tell you that we also have SASS FILE API support. Enable it for better result.   How this work This API feature are activated on 2 basis. 1. Feature are disabled by default on site (or you can change it if it’s not) 2. This API allow specific site to upload the files. Files upload may have some rule. For example (minimum or maximum size of file to uploaded, which format the site allowed you to upload). In case of resume site you will be allowed to use .doc (according to code of site)   How browser recognize that Site have SASS service. In HTML source of  the site, the code have a meta tag similar to this <meta name=”sass-upload-api” path=”/upload.json”/> Remember that upload.json is one file that has define the value of many settings {   "cookie_name": "ck_file",   "maximum_allowed_perday": 24,   "allowed_file_extensions","*.png,*.jpg,*.jpeg,*.gif",   "method": [       {           "get": "file/get",           "routing":"/file/get/{fileName}"       },       {           "post": "file/post",           "routing":"/file/post/{fileName}"       },       {           "delete": "file/delete",           "routing":"/file/delete/{fileName}"       },         {           "put": "file/put",           "routing":"/file/put/{fileName}"       },        {           "all": "file/all",           "routing":"/file/all/{fileName}"       }    ] } cookie name is simply a cookie which should be stored in browser and define in json. we define the cookie_name so we can easily share then with service in Windows system. This cookie will be accessible with the service so it’s security based safe. other cookie will not be shared.   The cookie will be post,put, get from this location. The all location will be simply about showing a whole list of file. This will gave a treeview kind of json to show the directories on sever.   for example example.com if you have activated the API with this site then you will seen a send to option in your explorer.exe when you send you will got a windows open which folder you want to use to send the file. The windows will also describe the limit and how much you can upload. This thing never required site to opened. When you upload the file this will be uploaded through FTP protocol. FTP protocol are better for performance.   How this API make thing faster. Suppose you want to ask a question and want to post image. you just do it and get it ready when you open stackoverflow.com now stackoverflow will only tell you which file you want to put on your current question that you asking for. second use is about people use cloud app.   There is no need of drag and drop anymore. we just need to do it without drag and drop it. we doesn’t need to open the site either. This thing is still in experiment level. I will update this post when I got some progress on this API.

    Read the article

  • Why do we use the Pythagorean theorem in game physics?

    - by Starkers
    I've recently learned that we use Pythagorean theorem a lot in our physics calculations and I'm afraid I don't really get the point. Here's an example from a book to make sure an object doesn't travel faster than a MAXIMUM_VELOCITY constant in the horizontal plane: MAXIMUM_VELOCITY = <any number>; SQUARED_MAXIMUM_VELOCITY = MAXIMUM_VELOCITY * MAXIMUM_VELOCITY; function animate(){ var squared_horizontal_velocity = (x_velocity * x_velocity) + (z_velocity * z_velocity); if( squared_horizontal_velocity <= SQUARED_MAXIMUM_VELOCITY ){ scalar = squared_horizontal_velocity / SQUARED_MAXIMUM_VELOCITY; x_velocity = x_velocity / scalar; z_velocity = x_velocity / scalar; } } Let's try this with some numbers: An object is attempting to move 5 units in x and 5 units in z. It should only be able to move 5 units horizontally in total! MAXIMUM_VELOCITY = 5; SQUARED_MAXIMUM_VELOCITY = 5 * 5; SQUARED_MAXIMUM_VELOCITY = 25; function animate(){ var x_velocity = 5; var z_velocity = 5; var squared_horizontal_velocity = (x_velocity * x_velocity) + (z_velocity * z_velocity); var squared_horizontal_velocity = 5 * 5 + 5 * 5; var squared_horizontal_velocity = 25 + 25; var squared_horizontal_velocity = 50; // if( squared_horizontal_velocity <= SQUARED_MAXIMUM_VELOCITY ){ if( 50 <= 25 ){ scalar = squared_horizontal_velocity / SQUARED_MAXIMUM_VELOCITY; scalar = 50 / 25; scalar = 2.0; x_velocity = x_velocity / scalar; x_velocity = 5 / 2.0; x_velocity = 2.5; z_velocity = z_velocity / scalar; z_velocity = 5 / 2.0; z_velocity = 2.5; // new_horizontal_velocity = x_velocity + z_velocity // new_horizontal_velocity = 2.5 + 2.5 // new_horizontal_velocity = 5 } } Now this works well, but we can do the same thing without Pythagoras: MAXIMUM_VELOCITY = 5; function animate(){ var x_velocity = 5; var z_velocity = 5; var horizontal_velocity = x_velocity + z_velocity; var horizontal_velocity = 5 + 5; var horizontal_velocity = 10; // if( horizontal_velocity >= MAXIMUM_VELOCITY ){ if( 10 >= 5 ){ scalar = horizontal_velocity / MAXIMUM_VELOCITY; scalar = 10 / 5; scalar = 2.0; x_velocity = x_velocity / scalar; x_velocity = 5 / 2.0; x_velocity = 2.5; z_velocity = z_velocity / scalar; z_velocity = 5 / 2.0; z_velocity = 2.5; // new_horizontal_velocity = x_velocity + z_velocity // new_horizontal_velocity = 2.5 + 2.5 // new_horizontal_velocity = 5 } } Benefits of doing it without Pythagoras: Less lines Within those lines, it's easier to read what's going on ...and it takes less time to compute, as there are less multiplications Seems to me like computers and humans get a better deal without Pythagorean theorem! However, I'm sure I'm wrong as I've seen Pythagoras' theorem in a number of reputable places, so I'd like someone to explain me the benefit of using Pythagorean theorem to a maths newbie. Does this have anything to do with unit vectors? To me a unit vector is when we normalize a vector and turn it into a fraction. We do this by dividing the vector by a larger constant. I'm not sure what constant it is. The total size of the graph? Anyway, because it's a fraction, I take it, a unit vector is basically a graph that can fit inside a 3D grid with the x-axis running from -1 to 1, z-axis running from -1 to 1, and the y-axis running from -1 to 1. That's literally everything I know about unit vectors... not much :P And I fail to see their usefulness. Also, we're not really creating a unit vector in the above examples. Should I be determining the scalar like this: // a mathematical work-around of my own invention. There may be a cleverer way to do this! I've also made up my own terms such as 'divisive_scalar' so don't bother googling var divisive_scalar = (squared_horizontal_velocity / SQUARED_MAXIMUM_VELOCITY); var divisive_scalar = ( 50 / 25 ); var divisive_scalar = 2; var multiplicative_scalar = (divisive_scalar / (2*divisive_scalar)); var multiplicative_scalar = (2 / (2*2)); var multiplicative_scalar = (2 / 4); var multiplicative_scalar = 0.5; x_velocity = x_velocity * multiplicative_scalar x_velocity = 5 * 0.5 x_velocity = 2.5 Again, I can't see why this is better, but it's more "unit-vector-y" because the multiplicative_scalar is a unit_vector? As you can see, I use words such as "unit-vector-y" so I'm really not a maths whiz! Also aware that unit vectors might have nothing to do with Pythagorean theorem so ignore all of this if I'm barking up the wrong tree. I'm a very visual person (3D modeller and concept artist by trade!) and I find diagrams and graphs really, really helpful so as many as humanely possible please!

    Read the article

  • Coherence Data Guarantees for Data Reads - Basic Terminology

    - by jpurdy
    When integrating Coherence into applications, each application has its own set of requirements with respect to data integrity guarantees. Developers often describe these requirements using expressions like "avoiding dirty reads" or "making sure that updates are transactional", but we often find that even in a small group of people, there may be a wide range of opinions as to what these terms mean. This may simply be due to a lack of familiarity, but given that Coherence sits at an intersection of several (mostly) unrelated fields, it may be a matter of conflicting vocabularies (e.g. "consistency" is similar but different in transaction processing versus multi-threaded programming). Since almost all data read consistency issues are related to the concept of concurrency, it is helpful to start with a definition of that, or rather what it means for two operations to be concurrent. Rather than implying that they occur "at the same time", concurrency is a slightly weaker statement -- it simply means that it can't be proven that one event precedes (or follows) the other. As an example, in a Coherence application, if two client members mutate two different cache entries sitting on two different cache servers at roughly the same time, it is likely that one update will precede the other by a significant amount of time (say 0.1ms). However, since there is no guarantee that all four members have their clocks perfectly synchronized, and there is no way to precisely measure the time it takes to send a given message between any two members (that have differing clocks), we consider these to be concurrent operations since we can not (easily) prove otherwise. So this leads to a question that we hear quite frequently: "Are the contents of the near cache always synchronized with the underlying distributed cache?". It's easy to see that if an update on a cache server results in a message being sent to each near cache, and then that near cache being updated that there is a window where the contents are different. However, this is irrelevant, since even if the application reads directly from the distributed cache, another thread update the cache before the read is returned to the application. Even if no other member modifies a cache entry prior to the local near cache entry being updated (and subsequently read), the purpose of reading a cache entry is to do something with the result, usually either displaying for consumption by a human, or by updating the entry based on the current state of the entry. In the former case, it's clear that if the data is updated faster than a human can perceive, then there is no problem (and in many cases this can be relaxed even further). For the latter case, the application must assume that the value might potentially be updated before it has a chance to update it. This almost aways the case with read-only caches, and the solution is the traditional optimistic transaction pattern, which requires the application to explicitly state what assumptions it made about the old value of the cache entry. If the application doesn't want to bother stating those assumptions, it is free to lock the cache entry prior to reading it, ensuring that no other threads will mutate the entry, a pessimistic approach. The optimistic approach relies on what is sometimes called a "fuzzy read". In other words, the application assumes that the read should be correct, but it also acknowledges that it might not be. (I use the qualifier "sometimes" because in some writings, "fuzzy read" indicates the situation where the application actually sees an original value and then later sees an updated value within the same transaction -- however, both definitions are roughly equivalent from an application design perspective). If the read is not correct it is called a "stale read". Going back to the definition of concurrency, it may seem difficult to precisely define a stale read, but the practical way of detecting a stale read is that is will cause the encompassing transaction to roll back if it tries to update that value. The pessimistic approach relies on a "coherent read", a guarantee that the value returned is not only the same as the primary copy of that value, but also that it will remain that way. In most cases this can be used interchangeably with "repeatable read" (though that term has additional implications when used in the context of a database system). In none of cases above is it possible for the application to perform a "dirty read". A dirty read occurs when the application reads a piece of data that was never committed. In practice the only way this can occur is with multi-phase updates such as transactions, where a value may be temporarily update but then withdrawn when a transaction is rolled back. If another thread sees that value prior to the rollback, it is a dirty read. If an application uses optimistic transactions, dirty reads will merely result in a lack of forward progress (this is actually one of the main risks of dirty reads -- they can be chained and potentially cause cascading rollbacks). The concepts of dirty reads, fuzzy reads, stale reads and coherent reads are able to describe the vast majority of requirements that we see in the field. However, the important thing is to define the terms used to define requirements. A quick web search for each of the terms in this article will show multiple meanings, so I've selected what are generally the most common variations, but it never hurts to state each definition explicitly if they are critical to the success of a project (many applications have sufficiently loose requirements that precise terminology can be avoided).

    Read the article

  • Some Early Considerations

    - by Chris Massey
    Following on from my previous post, I want to say "thank you" to everyone who has got in touch and got involved – you are pioneers! An update on where we are right now: paper prototypes v1 To be more specific, we’ve picked two of the ideas that seem to have more pros than cons, turned them into Balsamiq mockups, and are getting them fleshed out with realistic content. We’ll initially make these available to the aforementioned pioneers (thank you again), roll in the feedback, and then open up to get more data on what works and what doesn’t. If you’ve got any questions about this (or what we’re working on right now), feel free to ask me in the comments below. I’ve had a few people express an interest in the process we’re going through, and I’m more than happy to share details more frequently as we go along – not least because you, dear reader, will help us stay on target and create something Good. To start with, here’s a quick flashback to bring you all up to speed. A Brief Retrospective As you may already know, we’re creating a new publishing asset specifically focused on providing great content for web developers. We don’t yet know exactly what this thing will look like, or exactly how it will work, but we know we want to create something that is useful different. For my part, I’m seriously excited at the prospect of building a genuinely digital publishing system (as opposed to what most publishing is these days, which is print-style publishing which just happens to be on the web). The main challenge at this point is working out our build-measure-assess loop to speed up our experimental turn-around, and that’ll get better as we run more trials. Of course, there are a few things we’ve been pondering at this early conceptual stage: Do we publishing about heterogeneous technology stacks from day 1, or do we start with ASP.NET (which we’re familiar with) & branch out later? There are challenges with either approach. What publishing "modes" are already being well-handled? For example, the likes of Pluralsight, TekPub, and Treehouse have pretty much nailed video training (debate about price, if you like), and unless we think we can do it faster / better / cheaper (unlikely, for the record), we should leave them to it. Where should we base whatever we create? Should we create a completely new asset under a new name, graft something onto Simple-Talk (like the labs), or just build something directly into Simple-Talk? It sounds trivial, but it does have at least some impact on infrastructure and what how we manage the different types of content we (will) have. Are there any obvious problems or niches that we think could address really well, or should we just throw ideas out and see what readers respond to? What kind of users do we want to provide for? This actually deserves a little bit of unpacking… Why are you here? We currently divide readers into (broadly) the categories: Category 1: I know nothing about X, and I’d like to learn about it. Category 2: I know something about X, but I’d like to learn how to do something specific with it. Category 3: Ah man, I have a problem with X, and I need to fix it now. Now that I think about it, I might also include a 4th class of reader: Category 4: I’m looking for something interesting to engage my brain. These are clearly task-based categorizations, and depending on which task you’re performing when you arrive here, you’re going to need different types of content, or will have specific discovery needs. One of the questions that’s at the back of my mind whenever I consider a new idea is “How many of the categories will this satisfy?” As an example, typical video training is very well suited to categories 1, 2, and 4. StackOverflow is very well suited to category 3, and serves as a sign-posting system to the rest. Clearly it’s not necessary to satisfy every category need to be useful and popular, but being aware of what behavior readers might be exhibiting when they arrive will help us tune our ideas appropriately. < / Flashback > We don’t have clean answers to most of these considerations – they’re things we’re aware of, and each idea we look at is going to be best suited to a different mix of the options I’ve described. Our first experimental loop will be coming full circle in the next few days, so we should start to see how the different possibilities vary between ideas. Free to chime in with questions and suggestions about anything I’ve just brain-dumped, or at any stage as we go along. If you see anything that intrigued or enrages you, or just have an idea you’d like to share, I’d love to hear from you.

    Read the article

  • The Dreaded Startup Repair Loop on Win 7

    - by HighAltitudeCoder
    For most people, upgrading to Windows 7 has been a relatively painless process.  Not me.  I am in the unlucky 1% or less who had a somewhat less pleasant experience.  First, I cloned my entire onto a larger (and much faster) solid state hard drive, only experiencing minimal problems. Then, I bought the Retail version of Windows 7 Ultimate, took a deep breath and... oh yeah, I almost forgot - BACK UP THE COMPUTER.  The next morning I upgraded to Win 7 and everything seemed fine, until... I rebooted the system, the nice Windows 7 launch graphics come up, it's about to launch and AWWW, are you kidding me?!?!  Back to the BIOS splash screen?  Next comes the sequence of failure - attempt repair - unable to repair - do you want to wipe your hard drive decisions. Because I purchased the retail version, a number is provided where I could call Microsoft Tech support.  When I did, they instructed me to click "Install" from my installation CD, which did not work.  When I tried the "Upgrade" option, it reaches an impasse, telling you that yoiu have a newer version of Win 7, and thus cannot Upgrade.  If you choose "Install" you willl lose everything... files, programs, EVERYTHING.  Or at least this is what it tells you.  I was not willing to take the risk. To make things worse, I had installed a new antivirus software application before I realized my system was unstable (Trend Micro Titanium Internet Security), and this was causing additional problems. One interesting thing, and the only saving grace as it turns out, was that my system WOULD successfully reboot into the OS if I chose to restart it, rather than shut it down.  If I chose to shut down, I would have to go through the loop again until I was given the option to restart. As it turned out, I needed to update my BIOS.  I assumed that since I had updated my BIOS a long time ago to settings that were stable under Windows Vista Ultimate x64, I incorrectly expected Win 7 to adopt the same settings and didn't expect there to be any problems.  WRONG. My BIOS had a setting to halt the boot cycle if various kinds of errors were detected.  Windows Vista didn't care about this, but forget it under Windows 7.  I turned immediately corrected that BIOS setting.  Next, there were the two separate BIOS settings: enable USB mouse and enable USB keyboard.  The only sequence of events that would work were to start my reboot process over from stratch with a hard-wired non-usb keyboard and mouse.  Whent the system booted under these settings, it doesn't detect any errors due to either the mouse or keyboard, and actually booted for the first time in a long while (let me tell ya, that's an amazing experience after fiddling with settings for two entire weekends!) Next step: leave your old mouse and keyboard connected, but also connect your other two devices (mouse, keyboard) that use USB connections.  During the boot cycle, the operating system will not fail due to missing requirements during startup, and it will then pick up the new drivers necessary to use your new hardware. If you think you are in the clear here, you are wrong.  The next VERY IMPORTANT step is to remember to change your settings in the BIOS upon next startup.  Specifically, yoiu will need ot change your BIOS to enable USB mouse and enable USB keyboard input.  If you don't, Windows will detect an incompatibility upon the next startup, and you will be stuck once again in the endless cycle of reboot/Startup Repair/reboot/Startup Repair, without ever reaching a successful boot. Here's the thing - the BIOS and the drivers registered in Win 7 need to match.  If they don't, you're going to lose another weekend worrying and fiddling, all the while wondering if you've permanently damaged your hard drive beyond repair. (Sigh).  In the end, things worked out.  I must note that it is saddening to see how many posts there are out there that recommend just doing a clean install, as if it's the only option.  How many countless poor souls have lost their data, their backups, their pictures and videos, all for nothing other than the fact that the person giving advice just didn't know what to do at that point? My advice to you, try having a look at your BIOS settings first and making sure Win 7 can find your BIOS settings, and also disabling in your BIOS anything that might halt your system boot-up process if it encounters errors.

    Read the article

  • What's New in Oracle VM VirtualBox 4.2?

    - by Fat Bloke
    A year is a long time in the IT industry. Since the last VirtualBox feature release, which was a little over a year ago, we've seen: new releases of cool new operating systems, such as Windows 8, ChromeOS, and Mountain Lion; we've seen a myriad of new Linux releases from big Enterprise class distributions like Oracle 6.3, to accessible desktop distros like Ubuntu 12.04 and Fedora 17; and we've also seen the spec of a typical PC or laptop double in power. All of these events have influenced our new VirtualBox version which we're releasing today. Here's how... Powerful hosts  One of the trends we've seen is that as the average host platform becomes more powerful, our users are consistently running more and more vm's. Some of our users have large libraries of vm's of various vintages, whilst others have groups of vm's that are run together as an assembly of the various tiers in a multi-tiered software solution, for example, a database tier, middleware tier, and front-ends.  So we're pleased to unveil a more powerful VirtualBox Manager to address the needs of these users: VM Groups Groups allow you to organize your VM library in a sensible way, e.g.  by platform type, by project, by version, by whatever. To create groups you can drag one VM onto another or select one or more VM's and choose Machine...Group from the menu bar. You can expand and collapse groups to save screen real estate, and you can Enter and Leave a group (think iPad navigation here) by using the right and left arrow keys when groups are selected. But groups are more than passive folders, because you can now also perform operations on groups, rather than all the individual VMs. So if you have a multi-tiered solution you can start the whole stack up with just one click. Autostart Many VirtualBox users run dedicated services in their VMs, for example, running a Wiki. With these types of VM workloads, you really want the VM start up when the host machine boots up. So with 4.2 we've introduced a cross-platform Auto-start mechanism to allow you to treat VMs as host services. Headless VM Launching With VM's such as web servers, wikis, and other types of server-class workloads, the Console of the VM is pretty much redundant. For some time now VirtualBox has offered a separate launch mechanism for these VM's, namely the command-line interface commands VBoxHeadless or VBoxManage startvm ... --type headless commands. But with 4.2 we also allow you launch headless VMs from the Manager. Simply hold down Shift when launching the VM from the Manager.  It's that easy. But how do you stop a headless VM? Well, with 4.2 we allow you to Close the VM from the Manager. (BTW best to use the ACPI Shutdown method which allows the guest VM to close down gracefully.) Easy VM Creation For our expert users, the  New VM Wizard was a little tiresome, so now there's a faster 2-click VM creation mode. Just Hide the description when creating a new VM. Powerful VMs  As the hosts have become more powerful, so are the guests that are running inside them. Here are some of the 4.2 features to accommodate them: Virtual Network Interface Cards  With 4.2, it's now possible to create VMs with up to 36 NICs, when using the ICH9 chipset emulation. But with great power comes great responsibility (didn't Obi-Wan say something similar?), and so we have also introduced bandwidth limiting to prevent a rogue VM stealing the whole pipe. VLAN tagging Some of our users leverage VLANs extensively so we've enhanced the E1000 NICs to support this.  Processor Performance If you are running a CPU which supports Nested Paging (aka EPT in the Intel world) such as most of the Core i5 and i7 CPUs, or are running an AMD Bulldozer or later, you should see some performance improvements from our work with these processors. And while we're talking Processors, we've added support for some of the more modern VIA CPUs too. Powerful Automation Because VirtualBox runs atop a fully blown operating system, it makes sense to leverage the capabilities of the host to run scripts that can drive the guest VMs. Guest Automation was introduced in a prior release but with 4.2 we've revamped the APIs to allow a richer and more powerful set of operations to be executed by the guest. Check out the IGuest APIs in the VirtualBox Programming Guide and Reference (SDK). Powerful Platforms  All the hardcore engineering that has gone into 4.2 has been done for a purpose and that is to deliver a fast and powerful engine that can run almost any x86 OS because of the integrity of the virtualization. So we're pleased to add support for these platforms: Mac OS X "Mountain Lion"  Windows 8 Windows Server 2012 Ubuntu 12.04 (“Precise Pangolin”) Fedora 17 Oracle Linux 6.3  Here's the proof: We don't have time to go into the myriad of smaller improvements such as support for burning audio CDs from a guest, bi-directional clipboard control,  drag-and-drop of files into Linux guests, etc. so we'll leave that as an exercise for the user as soon as you've downloaded from the Oracle or community site and taken a peek at the User Guide. So all in all, a pretty solid release, one that we hope you'll enjoy discovering. - FB 

    Read the article

  • What's New in SGD 5.1?

    - by Fat Bloke
    Oracle announced the latest version of Secure Global Desktop (SGD) this week with 3 major themes: Support for Android devices; Support for Desktop Chrome clients;  Support for Oracle Unified Directory. I'll talk about the new features in a moment, but a bit of context first: Oracle SGD - what, how and why?  Oracle Secure Global Desktop is Oracle's secure remote access product which allows users on almost any device, to access almost any type application which  is hosted in the data center, from almost any location. And it does this by sitting on the edge of the datacenter, between the user and the applications: This is actually a really smart environment for an increasing number of use cases where: Users need mobility of location AND device (i.e. work from anywhere); IT needs to ensure security of applications and data (of course!) The application requires an end-user environment which can't be guaranteed and IT may not own the client platform (e.g. BYOD, working from home, partners or contractors). Oracle has a a specific interest in this of course. As the leading supplier of enterprise applications, many of Oracle's customers, and indeed Oracle itself, fit these criteria. So, as an IT guy rolling out an application to your employees, if one of your apps absolutely needs, say,  IE10 with Java 6 update 32, how can you be sure that the user population has this, especially when they're using their own devices? In the SGD model you, the IT guy, can set up, say, a Windows Server running the exact environment required, and then use SGD to publish this app, without needing to worry any further about the device the end user is using. What's new?  So back to SGD 5.1 and what is new there: Android devices Since we introduced our support for iPad tablets in SGD 5.0 we've had a big demand from customers to extend this to Android tablets too, and so we're pleased to announce that 5.1 supports Android 4.x tablets such as Nexus 7 and 10, and the Galaxy Tab. Here's how it works, with screenshots from my Nexus 7: Simply point your browser to the SGD server URL and login; The workspace is the list of apps that the admin has deemed ok for you to run. You click on an application to run it (here's Excel and Oracle E-Business Suite): There's an extended on-screen keyboard (extended because desktop apps need keys that don't appear on a tablet keyboard such as ctrl, WIndow key, etc) and touch gestures can be mapped to desktop events (such as tap and hold to right click) All in all a pretty nice implementation for Android tablet users. Desktop Chrome Browsers SGD has always been designed around using a browser to access your applications. But traditionally, this has involved using Java to deliver the SGD client component. With HTML5 and Javascript engines becoming so powerful, we thought we'd see how well a pure web client could perform with desktop apps. And the answer was, surprisingly well. So with this release we now offer this additional way of working, which can be enabled by a simple bit of configuration. Here's a Linux desktop running in a tab in Chrome. And if you resize the browser window, the Linux desktop is resized by SGD too. Very cool! Oracle Unified Directory As I mentioned above, a lot of Oracle users already benefit from SGD. And a lot of Oracle customers use Oracle Unified Directory as their Enterprise and Carrier grade user directory. So it makes a lot of sense that SGD now supports this LDAP directory for both Authentication and as a means to determine which users get which applications, e.g. publish the engineering app to the guys in the Development group, but give everyone E-Business Suite to let them do their expenses. Summary With new devices, and faster 4G networking becoming more prevalent, the pressure for businesses to move to a increasingly mobile enterprise is stronger than ever. SGD is good for users, and even better for IT. By offering the user the ability to work from anywhere, and IT the control and security they need, everyone wins with SGD. To try this for yourself, download SGD 5.1 (look under Desktop Virtualization Products) from the Oracle Software Delivery Cloud or if you're an existing customer, get it from My Oracle Support.  -FB 

    Read the article

  • What free space thresholds/limits are advisable for 640 GB and 2 TB hard disk drives with ZEVO ZFS on OS X?

    - by Graham Perrin
    Assuming that free space advice for ZEVO will not differ from advice for other modern implementations of ZFS … Question Please, what percentages or amounts of free space are advisable for hard disk drives of the following sizes? 640 GB 2 TB Thoughts A standard answer for modern implementations of ZFS might be "no more than 96 percent full". However if apply that to (say) a single-disk 640 GB dataset where some of the files most commonly used (by VirtualBox) are larger than 15 GB each, then I guess that blocks for those files will become sub optimally spread across the platters with around 26 GB free. I read that in most cases, fragmentation and defragmentation should not be a concern with ZFS. Sill, I like the mental picture of most fragments of a large .vdi in reasonably close proximity to each other. (Do features of ZFS make that wish for proximity too old-fashioned?) Side note: there might arise the question of how to optimise performance after a threshold is 'broken'. If it arises, I'll keep it separate. Background On a 640 GB StoreJet Transcend (product ID 0x2329) in the past I probably went beyond an advisable threshold. Currently the largest file is around 17 GB –  – and I doubt that any .vdi or other file on this disk will grow beyond 40 GB. (Ignore the purple masses, those are bundles of 8 MB band files.) Without HFS Plus: the thresholds of twenty, ten and five percent that I associate with Mobile Time Machine file system need not apply. I currently use ZEVO Community Edition 1.1.1 with Mountain Lion, OS X 10.8.2, but I'd like answers to be not too version-specific. References, chronological order ZFS Block Allocation (Jeff Bonwick's Blog) (2006-11-04) Space Maps (Jeff Bonwick's Blog) (2007-09-13) Doubling Exchange Performance (Bizarre ! Vous avez dit Bizarre ?) (2010-03-11) … So to solve this problem, what went in 2010/Q1 software release is multifold. The most important thing is: we increased the threshold at which we switched from 'first fit' (go fast) to 'best fit' (pack tight) from 70% full to 96% full. With TB drives, each slab is at least 5GB and 4% is still 200MB plenty of space and no need to do anything radical before that. This gave us the biggest bang. Second, instead of trying to reuse the same primary slabs until it failed an allocation we decided to stop giving the primary slab this preferential threatment as soon as the biggest allocation that could be satisfied by a slab was down to 128K (metaslab_df_alloc_threshold). At that point we were ready to switch to another slab that had more free space. We also decided to reduce the SMO bonus. Before, a slab that was 50% empty was preferred over slabs that had never been used. In order to foster more write aggregation, we reduced the threshold to 33% empty. This means that a random write workload now spread to more slabs where each one will have larger amount of free space leading to more write aggregation. Finally we also saw that slab loading was contributing to lower performance and implemented a slab prefetch mechanism to reduce down time associated with that operation. The conjunction of all these changes lead to 50% improved OLTP and 70% reduced variability from run to run … OLTP Improvements in Sun Storage 7000 2010.Q1 (Performance Profiles) (2010-03-11) Alasdair on Everything » ZFS runs really slowly when free disk usage goes above 80% (2010-07-18) where commentary includes: … OpenSolaris has changed this in onnv revision 11146 … [CFT] Improved ZFS metaslab code (faster write speed) (2010-08-22)

    Read the article

  • ERROR 2003 (HY000): Can't connect to MySQL server on (111)

    - by JohnMerlino
    I am unable to connect to on my ubuntu installation a remote tcp/ip which contains a mysql installation: viggy@ubuntu:~$ mysql -u user.name -p -h xxx.xxx.xxx.xxx -P 3306 Enter password: ERROR 2003 (HY000): Can't connect to MySQL server on 'xxx.xxx.xxx.xxx' (111) I commented out the line below using vim in /etc/mysql/my.cnf: # Instead of skip-networking the default is now to listen only on # localhost which is more compatible and is not less secure. #bind-address = 127.0.0.1 Then I restarted the server: sudo service mysql restart But still I get the same error. This is the content of my.cnf: # # The MySQL database server configuration file. # # You can copy this to one of: # - "/etc/mysql/my.cnf" to set global options, # - "~/.my.cnf" to set user-specific options. # # One can use all long options that the program supports. # Run program with --help to get a list of available options and with # --print-defaults to see which it would actually understand and use. # # For explanations see # http://dev.mysql.com/doc/mysql/en/server-system-variables.html # This will be passed to all mysql clients # It has been reported that passwords should be enclosed with ticks/quotes # escpecially if they contain "#" chars... # Remember to edit /etc/mysql/debian.cnf when changing the socket location. [client] port = 3306 socket = /var/run/mysqld/mysqld.sock # Here is entries for some specific programs # The following values assume you have at least 32M ram # This was formally known as [safe_mysqld]. Both versions are currently parsed. [mysqld_safe] socket = /var/run/mysqld/mysqld.sock nice = 0 [mysqld] # # * Basic Settings # user = mysql pid-file = /var/run/mysqld/mysqld.pid socket = /var/run/mysqld/mysqld.sock port = 3306 basedir = /usr datadir = /var/lib/mysql tmpdir = /tmp lc-messages-dir = /usr/share/mysql skip-external-locking # # Instead of skip-networking the default is now to listen only on # localhost which is more compatible and is not less secure. #bind-address = 127.0.0.1 # # * Fine Tuning # key_buffer = 16M max_allowed_packet = 16M thread_stack = 192K thread_cache_size = 8 # This replaces the startup script and checks MyISAM tables if needed # the first time they are touched myisam-recover = BACKUP #max_connections = 100 #table_cache = 64 #thread_concurrency = 10 # # * Query Cache Configuration # query_cache_limit = 1M query_cache_size = 16M # # * Logging and Replication # # Both location gets rotated by the cronjob. # Be aware that this log type is a performance killer. # As of 5.1 you can enable the log at runtime! #general_log_file = /var/log/mysql/mysql.log #general_log = 1 # # Error logging goes to syslog due to /etc/mysql/conf.d/mysqld_safe_syslog.cnf. # # Here you can see queries with especially long duration #log_slow_queries = /var/log/mysql/mysql-slow.log #long_query_time = 2 #log-queries-not-using-indexes # # The following can be used as easy to replay backup logs or for replication. # note: if you are setting up a replication slave, see README.Debian about # other settings you may need to change. #server-id = 1 #log_bin = /var/log/mysql/mysql-bin.log expire_logs_days = 10 max_binlog_size = 100M #binlog_do_db = include_database_name #binlog_ignore_db = include_database_name # # * InnoDB # # InnoDB is enabled by default with a 10MB datafile in /var/lib/mysql/. # Read the manual for more InnoDB related options. There are many! # # * Security Features # # Read the manual, too, if you want chroot! # chroot = /var/lib/mysql/ # # For generating SSL certificates I recommend the OpenSSL GUI "tinyca". # # ssl-ca=/etc/mysql/cacert.pem # ssl-cert=/etc/mysql/server-cert.pem # ssl-key=/etc/mysql/server-key.pem [mysqldump] quick quote-names max_allowed_packet = 16M [mysql] #no-auto-rehash # faster start of mysql but no tab completition [isamchk] key_buffer = 16M # # * IMPORTANT: Additional settings that can override those from this file! # The files must end with '.cnf', otherwise they'll be ignored. # !includedir /etc/mysql/conf.d/ (Note that I can log into my local mysql install just fine by running mysql (and it will log me in as root) and also note that I can get into mysql in the remote server by logging into via ssh and then invoking mysql), but I am unable to connect to the remote server via my terminal using the host, and I need to do it that way so that I can then use mysql workbench.

    Read the article

  • High Load mysql on Debian server

    - by Oleg Abrazhaev
    I have Debian server with 32 gb memory. And there is apache2, memcached and nginx on this server. Memory load always on maximum. Only 500m free. Most memory leak do MySql. Apache only 70 clients configured, other services small memory usage. When mysql use all memory it stops. And nothing works, need mysql reboot. Mysql configured use maximum 24 gb memory. I have hight weight InnoDB bases. (400000 rows, 30 gb). And on server multithread daemon, that makes many inserts in this tables, thats why InnoDB. There is my mysql config. [mysqld] # # * Basic Settings # default-time-zone = "+04:00" user = mysql pid-file = /var/run/mysqld/mysqld.pid socket = /var/run/mysqld/mysqld.sock port = 3306 basedir = /usr datadir = /var/lib/mysql tmpdir = /tmp language = /usr/share/mysql/english skip-external-locking default-time-zone='Europe/Moscow' # # Instead of skip-networking the default is now to listen only on # localhost which is more compatible and is not less secure. # # * Fine Tuning # #low_priority_updates = 1 concurrent_insert = ALWAYS wait_timeout = 600 interactive_timeout = 600 #normal key_buffer_size = 2024M #key_buffer_size = 1512M #70% hot cache key_cache_division_limit= 70 #16-32 max_allowed_packet = 32M #1-16M thread_stack = 8M #40-50 thread_cache_size = 50 #orderby groupby sort sort_buffer_size = 64M #same myisam_sort_buffer_size = 400M #temp table creates when group_by tmp_table_size = 3000M #tables in memory max_heap_table_size = 3000M #on disk open_files_limit = 10000 table_cache = 10000 join_buffer_size = 5M # This replaces the startup script and checks MyISAM tables if needed # the first time they are touched myisam-recover = BACKUP #myisam_use_mmap = 1 max_connections = 200 thread_concurrency = 8 # # * Query Cache Configuration # #more ignored query_cache_limit = 50M query_cache_size = 210M #on query cache query_cache_type = 1 # # * Logging and Replication # # Both location gets rotated by the cronjob. # Be aware that this log type is a performance killer. #log = /var/log/mysql/mysql.log # # Error logging goes to syslog. This is a Debian improvement :) # # Here you can see queries with especially long duration log_slow_queries = /var/log/mysql/mysql-slow.log long_query_time = 1 log-queries-not-using-indexes # # The following can be used as easy to replay backup logs or for replication. # note: if you are setting up a replication slave, see README.Debian about # other settings you may need to change. #server-id = 1 #log_bin = /var/log/mysql/mysql-bin.log server-id = 1 log-bin = /var/lib/mysql/mysql-bin #replicate-do-db = gate log-bin-index = /var/lib/mysql/mysql-bin.index log-error = /var/lib/mysql/mysql-bin.err relay-log = /var/lib/mysql/relay-bin relay-log-info-file = /var/lib/mysql/relay-bin.info relay-log-index = /var/lib/mysql/relay-bin.index binlog_do_db = 24avia expire_logs_days = 10 max_binlog_size = 100M read_buffer_size = 4024288 innodb_buffer_pool_size = 5000M innodb_flush_log_at_trx_commit = 2 innodb_thread_concurrency = 8 table_definition_cache = 2000 group_concat_max_len = 16M #binlog_do_db = gate #binlog_ignore_db = include_database_name # # * BerkeleyDB # # Using BerkeleyDB is now discouraged as its support will cease in 5.1.12. #skip-bdb # # * InnoDB # # InnoDB is enabled by default with a 10MB datafile in /var/lib/mysql/. # Read the manual for more InnoDB related options. There are many! # You might want to disable InnoDB to shrink the mysqld process by circa 100MB. #skip-innodb # # * Security Features # # Read the manual, too, if you want chroot! # chroot = /var/lib/mysql/ # # For generating SSL certificates I recommend the OpenSSL GUI "tinyca". # # ssl-ca=/etc/mysql/cacert.pem # ssl-cert=/etc/mysql/server-cert.pem # ssl-key=/etc/mysql/server-key.pem [mysqldump] quick quote-names max_allowed_packet = 500M [mysql] #no-auto-rehash # faster start of mysql but no tab completition [isamchk] key_buffer = 32M key_buffer_size = 512M # # * NDB Cluster # # See /usr/share/doc/mysql-server-*/README.Debian for more information. # # The following configuration is read by the NDB Data Nodes (ndbd processes) # not from the NDB Management Nodes (ndb_mgmd processes). # # [MYSQL_CLUSTER] # ndb-connectstring=127.0.0.1 # # * IMPORTANT: Additional settings that can override those from this file! # The files must end with '.cnf', otherwise they'll be ignored. # !includedir /etc/mysql/conf.d/ Please, help me make it stable. Memory used /etc/mysql # free total used free shared buffers cached Mem: 32930800 32766424 164376 0 139208 23829196 -/+ buffers/cache: 8798020 24132780 Swap: 33553328 44660 33508668 Maybe my problem not in memory, but MySQL stops every day. As you can see, cache memory free 24 gb. Thank to Michael Hampton? for correction. Load overage on server 3.5. Maybe hdd or another problem? Maybe my config not optimal for 30gb InnoDB ?

    Read the article

  • Exchange Server 2007 Setup

    - by AlamedaDad
    Hi, I'm working on a upgrade to Exchange 2007 and I wanted to get some advise on hardware choices. We currently have an Exchange 2003 STD server with 400 users split between 6 AD Sites, that is housed on a single server. We need to move to a redundant, fault tolerant system to support our users. I'm planning on installing 2 Dell 1950 servers with W2k8-std to act as CAS and Hub servers, with NLB to allow abstraction of the actual server name to the users. There won't be an edge system since we have a Barracuda box already that will handle in/out spam/virus filtering. Backend I'm planning on 2 mailbox servers which will be Dell 2950s with 16GB RAM, 2 either dual-core or quad-core CPUs and 6 300GB SAS drives in some RAID config. These systems will be clustered using W2k8 Ent clustering and running CCR in Exchange. My questions are as follows: Is 16GB enough RAM for serving that many mailboxes along with the windows clustering and ccr? I'm trying to figure out disk layouts and I'm unsure of whether to use all local disk or some local and some SAN, via an OpenFiler iSCSI server. The SAN would be a Dell 2850 with 6 - 300GB SCSI drives and a PERC controller to slice as I want, with 8GB RAM. Option 1: 2 drives, RAID 1 - OS 2 drives, RAID 1 - Logs 2 drives, RAID 1 - Mail stores Option 2: 2 drives, RAID 1 - OS and logs 4 drives, RAID 5 - Mail Stores and scratch space for eseutil. Option 3: 2 drives, RAID 1 - OS 2 drives, RAID 1 - Logs 2 drives, RAID 0 - scratch space ~300GB iSCSI volume for mail stores Option 4: 2 drives, RAID 1 - OS 4 drives, RAID 5 - scratch space ~300GB iSCSI volume for mail stores ~300GB iSCSI volume for logs I have 2 sockets for CPUs and need to chose between dual and quad cores. The dual core have faster clocks but less cache and I'm thinking older architecture. Am I better off with more cores and cache while sacraficing clock speed? I am planning on adding the new E2K7 cluster to the E2K3 server and then move each mailbox over, all at once, then remove the old server. This seems more complicated than simply getting rid of the 2003 server and then adding the 2007 cluster and restoring the mailboxes using PowerControls or exmerge. The migration option lets me do this on my time, where a cutover means it all needs to work at once. If I go with the cutover method, how can I prebuild the servers and add them to the domain right after removing the 2003 server, or can't I? I think the answer is no and the migration is my only real option if I want to prebuild. I need to also migrate about 30GB of Public Folders. Is there anything special about this, other than specifying in the E2K7 install that I want older Outlook clients and PF's setup? I guess I could even keep the E2K3 server to host just the PFs? Lastly, if I have a mix of Outlook 200, 2003 and 2007 what do I need to do to make sure they all have access to the GAL and OAB? At time of cutover, we'll be at like 90% 2007, but we will have some older stuff around. My plan is to use Outlook Anywhere on laptops that are used outside the physical network. Are there any gotchas involved in that? I'm even thinking about using is for all Outlook clients, does anyone do that? The reason I'm considering it is that our WAN is really VPN tunnels over internet connections, so not a fully messhed, stable WAN. Thank you all very much for the assistance in advance and I look forward to discussion of these points! Regards...Michael

    Read the article

  • How do I get my ubuntu server to listen for database connections?

    - by Bob Flemming
    I am having a problems connecting to my database outside of phpmyadmin. Im pretty sure this is because my server isn't listening on port 3306. When I type: sudo netstat -ntlp on my OTHER working server I can see the following line: tcp 0 0 0.0.0.0:3306 0.0.0.0:* LISTEN 20445/mysqld However, this line does not appear on the server I am having difficulty with. How do I make my sever listen for mysql connections? Here my my.conf file: # # The MySQL database server configuration file. # # You can copy this to one of: # - "/etc/mysql/my.cnf" to set global options, # - "~/.my.cnf" to set user-specific options. # # One can use all long options that the program supports. # Run program with --help to get a list of available options and with # --print-defaults to see which it would actually understand and use. # # For explanations see # http://dev.mysql.com/doc/mysql/en/server-system-variables.html # This will be passed to all mysql clients # It has been reported that passwords should be enclosed with ticks/quotes # escpecially if they contain "#" chars... # Remember to edit /etc/mysql/debian.cnf when changing the socket location. [client] port = 3306 socket = /var/run/mysqld/mysqld.sock # Here is entries for some specific programs # The following values assume you have at least 32M ram # This was formally known as [safe_mysqld]. Both versions are currently parsed. [mysqld_safe] socket = /var/run/mysqld/mysqld.sock nice = 0 [mysqld] # # * Basic Settings # user = mysql pid-file = /var/run/mysqld/mysqld.pid socket = /var/run/mysqld/mysqld.sock port = 3306 basedir = /usr datadir = /var/lib/mysql tmpdir = /tmp lc-messages-dir = /usr/share/mysql #skip-networking=off #skip_networking=off #skip-external-locking # # Instead of skip-networking the default is now to listen only on # localhost which is more compatible and is not less secure. #bind-address = 0.0.0.0 # # * Fine Tuning # key_buffer = 64M max_allowed_packet = 64M thread_stack = 650K thread_cache_size = 32 # This replaces the startup script and checks MyISAM tables if needed # the first time they are touched myisam-recover = BACKUP #max_connections = 100 #table_cache = 64 #thread_concurrency = 10 # # * Query Cache Configuration # query_cache_limit = 2M query_cache_size = 32M # # * Logging and Replication # # Both location gets rotated by the cronjob. # Be aware that this log type is a performance killer. # As of 5.1 you can enable the log at runtime! #general_log_file = /var/log/mysql/mysql.log #general_log = 1 # # Error logging goes to syslog due to /etc/mysql/conf.d/mysqld_safe_syslog.cnf. # # Here you can see queries with especially long duration #log_slow_queries = /var/log/mysql/mysql-slow.log #long_query_time = 2 #log-queries-not-using-indexes # # The following can be used as easy to replay backup logs or for replication. # note: if you are setting up a replication slave, see README.Debian about # other settings you may need to change. #server-id = 1 #log_bin = /var/log/mysql/mysql-bin.log expire_logs_days = 10 max_binlog_size = 100M #binlog_do_db = include_database_name #binlog_ignore_db = include_database_name # # * InnoDB # # InnoDB is enabled by default with a 10MB datafile in /var/lib/mysql/. # Read the manual for more InnoDB related options. There are many! # # * Security Features # # Read the manual, too, if you want chroot! # chroot = /var/lib/mysql/ # # For generating SSL certificates I recommend the OpenSSL GUI "tinyca". # # ssl-ca=/etc/mysql/cacert.pem # ssl-cert=/etc/mysql/server-cert.pem # ssl-key=/etc/mysql/server-key.pem [mysqldump] quick quote-names max_allowed_packet = 32M [mysql] #no-auto-rehash # faster start of mysql but no tab completition [isamchk] key_buffer = 32M # # * IMPORTANT: Additional settings that can override those from this file! # The files must end with '.cnf', otherwise they'll be ignored. # !includedir /etc/mysql/conf.d/

    Read the article

  • AMD 24 core server memory bandwidth

    - by ntherning
    I need some help to determine whether the memory bandwidth I'm seeing under Linux on my server is normal or not. Here's the server spec: HP ProLiant DL165 G7 2x AMD Opteron 6164 HE 12-Core 40 GB RAM (10 x 4GB DDR1333) Debian 6.0 Using mbw on this server I get the following numbers: foo1:~# mbw -n 3 1024 Long uses 8 bytes. Allocating 2*134217728 elements = 2147483648 bytes of memory. Using 262144 bytes as blocks for memcpy block copy test. Getting down to business... Doing 3 runs per test. 0 Method: MEMCPY Elapsed: 0.58047 MiB: 1024.00000 Copy: 1764.082 MiB/s 1 Method: MEMCPY Elapsed: 0.58012 MiB: 1024.00000 Copy: 1765.152 MiB/s 2 Method: MEMCPY Elapsed: 0.58010 MiB: 1024.00000 Copy: 1765.201 MiB/s AVG Method: MEMCPY Elapsed: 0.58023 MiB: 1024.00000 Copy: 1764.811 MiB/s 0 Method: DUMB Elapsed: 0.36174 MiB: 1024.00000 Copy: 2830.778 MiB/s 1 Method: DUMB Elapsed: 0.35869 MiB: 1024.00000 Copy: 2854.817 MiB/s 2 Method: DUMB Elapsed: 0.35848 MiB: 1024.00000 Copy: 2856.481 MiB/s AVG Method: DUMB Elapsed: 0.35964 MiB: 1024.00000 Copy: 2847.310 MiB/s 0 Method: MCBLOCK Elapsed: 0.23546 MiB: 1024.00000 Copy: 4348.860 MiB/s 1 Method: MCBLOCK Elapsed: 0.23544 MiB: 1024.00000 Copy: 4349.230 MiB/s 2 Method: MCBLOCK Elapsed: 0.23544 MiB: 1024.00000 Copy: 4349.359 MiB/s AVG Method: MCBLOCK Elapsed: 0.23545 MiB: 1024.00000 Copy: 4349.149 MiB/s On one of my other servers (based on Intel Xeon E3-1270): foo2:~# mbw -n 3 1024 Long uses 8 bytes. Allocating 2*134217728 elements = 2147483648 bytes of memory. Using 262144 bytes as blocks for memcpy block copy test. Getting down to business... Doing 3 runs per test. 0 Method: MEMCPY Elapsed: 0.18960 MiB: 1024.00000 Copy: 5400.901 MiB/s 1 Method: MEMCPY Elapsed: 0.18922 MiB: 1024.00000 Copy: 5411.690 MiB/s 2 Method: MEMCPY Elapsed: 0.18944 MiB: 1024.00000 Copy: 5405.491 MiB/s AVG Method: MEMCPY Elapsed: 0.18942 MiB: 1024.00000 Copy: 5406.024 MiB/s 0 Method: DUMB Elapsed: 0.14838 MiB: 1024.00000 Copy: 6901.200 MiB/s 1 Method: DUMB Elapsed: 0.14818 MiB: 1024.00000 Copy: 6910.561 MiB/s 2 Method: DUMB Elapsed: 0.14820 MiB: 1024.00000 Copy: 6909.628 MiB/s AVG Method: DUMB Elapsed: 0.14825 MiB: 1024.00000 Copy: 6907.127 MiB/s 0 Method: MCBLOCK Elapsed: 0.04362 MiB: 1024.00000 Copy: 23477.623 MiB/s 1 Method: MCBLOCK Elapsed: 0.04262 MiB: 1024.00000 Copy: 24025.151 MiB/s 2 Method: MCBLOCK Elapsed: 0.04258 MiB: 1024.00000 Copy: 24048.849 MiB/s AVG Method: MCBLOCK Elapsed: 0.04294 MiB: 1024.00000 Copy: 23847.599 MiB/s For reference here's what I get on my Intel based laptop: laptop:~$ mbw -n 3 1024 Long uses 8 bytes. Allocating 2*134217728 elements = 2147483648 bytes of memory. Using 262144 bytes as blocks for memcpy block copy test. Getting down to business... Doing 3 runs per test. 0 Method: MEMCPY Elapsed: 0.40566 MiB: 1024.00000 Copy: 2524.269 MiB/s 1 Method: MEMCPY Elapsed: 0.38458 MiB: 1024.00000 Copy: 2662.638 MiB/s 2 Method: MEMCPY Elapsed: 0.38876 MiB: 1024.00000 Copy: 2634.043 MiB/s AVG Method: MEMCPY Elapsed: 0.39300 MiB: 1024.00000 Copy: 2605.600 MiB/s 0 Method: DUMB Elapsed: 0.30707 MiB: 1024.00000 Copy: 3334.745 MiB/s 1 Method: DUMB Elapsed: 0.30425 MiB: 1024.00000 Copy: 3365.653 MiB/s 2 Method: DUMB Elapsed: 0.30342 MiB: 1024.00000 Copy: 3374.849 MiB/s AVG Method: DUMB Elapsed: 0.30491 MiB: 1024.00000 Copy: 3358.328 MiB/s 0 Method: MCBLOCK Elapsed: 0.07875 MiB: 1024.00000 Copy: 13003.670 MiB/s 1 Method: MCBLOCK Elapsed: 0.08374 MiB: 1024.00000 Copy: 12228.034 MiB/s 2 Method: MCBLOCK Elapsed: 0.07635 MiB: 1024.00000 Copy: 13411.216 MiB/s AVG Method: MCBLOCK Elapsed: 0.07961 MiB: 1024.00000 Copy: 12862.006 MiB/s So according to mbw my laptop is 3 times faster than the server!!! Please help me explain this. I've also tried to mount a ram disk and use dd to benchmark it and I get similar differences so I don't think mbw is to blame. I've checked the BIOS settings and the memory seem to be running at full speed. According to the hosting company the modules are all OK. Could this have something to do with NUMA? It seems like Node Interleaving is disabled on this server. Will enabling it (thus turning off NUMA) make a difference? foo1:~# numactl --hardware available: 4 nodes (0-3) node 0 cpus: 0 1 2 3 4 5 node 0 size: 8190 MB node 0 free: 7898 MB node 1 cpus: 6 7 8 9 10 11 node 1 size: 12288 MB node 1 free: 12073 MB node 2 cpus: 18 19 20 21 22 23 node 2 size: 12288 MB node 2 free: 12034 MB node 3 cpus: 12 13 14 15 16 17 node 3 size: 8192 MB node 3 free: 8032 MB node distances: node 0 1 2 3 0: 10 20 20 20 1: 20 10 20 20 2: 20 20 10 20 3: 20 20 20 10

    Read the article

  • Linux not buffering block I/O when the device is not "in use" (i.e. mounted)

    - by Radek Hladík
    I am installing new server and I've found an interesting issue. The server is running Fedora 19 (3.11.7-200.fc19.x86_64 kernel) and is supposed to host a few KVM/Qemu virtual servers (mail server, file server, etc..). The HW is Intel(R) Xeon(R) CPU 5160 @ 3.00GHz with 16GB RAM. One of the most important features will be Samba server and we have decided to make it as virtual machine with almost direct access to the disks. So the real HDD is cached on SSD (via bcache) then raided with md and the final device is exported into the virtual machine via virtio. The virtual machine is again Fedora 19 with the same kernel. One important topic to find out is whether the virtualization layer will not introduce high overload into disk I/Os. So far I've been able to get up to 180MB/s in VM and up to 220MB/s on real HW (on the SSD disk). I am still not sure why the overhead is so big but it is more than the network can handle so I do not care so much. The interesting thing is that I've found that the disk reads are not buffered in the VM unless I create and mount FS on the disk or I use the disks somehow. Simply put: Lets do dd to read disk for the first time (the /dev/vdd is an old Raptor disk 70MB/s is its real speed): [root@localhost ~]# dd if=/dev/vdd of=/dev/null bs=256k count=10000 ; cat /proc/meminfo | grep Buffers 2621440000 bytes (2.6 GB) copied, 36.8038 s, 71.2 MB/s Buffers: 14444 kB Rereading the data shows that they are cached somewhere but not in buffers of the VM. Also the speed increased to "only" 500MB/s. The VM has 4GB of RAM (more that the test file) [root@localhost ~]# dd if=/dev/vdd of=/dev/null bs=256k count=10000 ; cat /proc/meminfo | grep Buffers 2621440000 bytes (2.6 GB) copied, 5.16016 s, 508 MB/s Buffers: 14444 kB [root@localhost ~]# dd if=/dev/vdd of=/dev/null bs=256k count=10000 ; cat /proc/meminfo | grep Buffers 2621440000 bytes (2.6 GB) copied, 5.05727 s, 518 MB/s Buffers: 14444 kB Now lets mount the FS on /dev/vdd and try the dd again: [root@localhost ~]# mount /dev/vdd /mnt/tmp [root@localhost ~]# dd if=/dev/vdd of=/dev/null bs=256k count=10000 ; cat /proc/meminfo | grep Buffers 2621440000 bytes (2.6 GB) copied, 4.68578 s, 559 MB/s Buffers: 2574592 kB [root@localhost ~]# dd if=/dev/vdd of=/dev/null bs=256k count=10000 ; cat /proc/meminfo | grep Buffers 2621440000 bytes (2.6 GB) copied, 1.50504 s, 1.7 GB/s Buffers: 2574592 kB While the first read was the same, all 2.6GB got buffered and the next read was at 1.7GB/s. And when I unmount the device: [root@localhost ~]# umount /mnt/tmp [root@localhost ~]# cat /proc/meminfo | grep Buffers Buffers: 14452 kB [root@localhost ~]# dd if=/dev/vdd of=/dev/null bs=256k count=10000 ; cat /proc/meminfo | grep Buffers 2621440000 bytes (2.6 GB) copied, 5.10499 s, 514 MB/s Buffers: 14468 kB The bcache was disabled while testing and the results are same on faster (newer) HDDs and on SSD (except for the initial read speed of course). To sum it up. When I read from the device via dd first time, it gets read from the disk. Next time I reread it gets cached in the host but not in the guest (thats actually the same issue, more on that later). When I mount the filesystem but try to read the device directly it gets cached in VM (via buffers). As soon as I stop "using" it, buffers are discarded and the device is not cached anymore in the VM. When I looked into buffers value on the host I realized that the situation is the same. The block I/O gets buffered only when the disk is in use, in this case it means "exported to a VM". On host, after all the measurement done: 3165552 buffers On the host, after the VM shutdown: 119176 buffers I know it is not important as the disks will be mounted all the time but I am curious and I would like to know why it is working like this.

    Read the article

  • Windows 7: How to place SuperFetch cache on an SSD?

    - by Ian Boyd
    I'm thinking of adding a solid state drive (SSD) to my existing Windows 7 installation. I know I can (and should) move my paging file to the SSD: Should the pagefile be placed on SSDs? Yes. Most pagefile operations are small random reads or larger sequential writes, both of which are types of operations that SSDs handle well. In looking at telemetry data from thousands of traces and focusing on pagefile reads and writes, we find that Pagefile.sys reads outnumber pagefile.sys writes by about 40 to 1, Pagefile.sys read sizes are typically quite small, with 67% less than or equal to 4 KB, and 88% less than 16 KB. Pagefile.sys writes are relatively large, with 62% greater than or equal to 128 KB and 45% being exactly 1 MB in size. In fact, given typical pagefile reference patterns and the favorable performance characteristics SSDs have on those patterns, there are few files better than the pagefile to place on an SSD. What I don't know is if I even can put a SuperFetch cache (i.e. ReadyBoost cache) on the solid state drive. I want to get the benefit of Windows being able to cache gigabytes of frequently accessed data on a relativly small (e.g. 30GB) solid state drive. This is exactly what SuperFetch+ReadyBoost (or SuperFetch+ReadyDrive) was designed for. Will Windows offer (or let) me place a ReadyBoost cache on a solid state flash drive connected via SATA? A problem with the ReadyBoost cache over the ReadyDrive cache is that the ReadyBoost cache does not survive between reboots. The cache is encrypted with a per-session key, making its existing contents unusable during boot and SuperFetch pre-fetching during login. Update One I know that Windows Vista limited you to only one ReadyBoost.sfcache file (I do not know if Windows 7 removed that limitation): Q: Can use use multiple devices for EMDs? A: Nope. We've limited Vista to one ReadyBoost per machine Q: Why just one device? A: Time and quality. Since this is the first revision of the feature, we decided to focus on making the single device exceptional, without the difficulties of managing multiple caches. We like the idea, though, and it's under consideration for future versions. I also know that the 4GB limit on the cache file was a limitation of the FAT filesystem used on most USB sticks - an SSD drive would be formatted with NTFS: Q: What's the largest amount of flash that I can use for ReadyBoost? A: You can use up to 4GB of flash for ReadyBoost (which turns out to be 8GB of cache w/ the compression) Q: Why can't I use more than 4GB of flash? A: The FAT32 filesystem limits our ReadyBoost.sfcache file to 4GB Can a ReadyBoost cache on an NTFS volume be larger than 4GB? Update Two The ReadyBoost cache is encrypted with a per-boot session key. This means that the cache has to be re-built after each boot, and cannot be used to help speed boot times, or latency from login to usable. Windows ReadyDrive technology takes advantage of non-volatile (NV) memory (i.e. flash) that is incorporated with some hybrid hard drives. This flash cache can be used to help Windows boot, or resume from hibernate faster. Will Windows 7 use an internal SSD drive as a ReadyBoost/*ReadyDrive*/SuperFetch cache? Is it possible to make Windows store a SuperFetch cache (i.e. ReadyBoost) on a non-removable SSD? Is it possible to not encrypt the ReadyBoost cache, and if so will Windows 7 use the cache at boot time? See also SuperUser.com: ReadyBoost + SSD = ? Windows 7 - ReadyBoost & SSD drives? Support and Q&A for Solid-State Drives Using SDD as a cache for HDD, is there a solution? Performance increase using SSD for paging/fetch/cache or ReadyBoost? (Win7) Windows 7 To Boost SSD Performance How to Disable Nonvolatile Caching

    Read the article

  • BluRay audio/video stuttering with PowerDVD 11, WinDVD 11 Pro, etc? Xonar/Auzen HD audio option?

    - by jrista
    I recently upgraded my Windows 7 MediaCenter HTPC due to a motherboard failure (really old motherboard and cpu, it was on its last legs.) I chose to upgrade to an i5 system with everything built into the motherboard. I did my due diligence, researched, and found some hardware that was within my budget. I ended up with: Core i5 2500K (3.3Ghz) Corsair XMS3 2x2Gb DDR3 (4Gb) ASUS P8H 61-M LE/CSM MicroCenter 64Gb SSD (Previous BluRay player, forget the brand) The system is pretty awesome, and plays everything I have perfectly. I almost went with an Atom solution, however there have been numerous notes that they do not play NetFlix Instant Watch well...and I am a heavy Netflix IW user. High definition BluRay rips work well, although they usually contain lower audio quality than the BluRay's they were ripped from. The real problem I am encountering is playing back BluRay video from discs. For some reason, I am encountering rather terrible stuttering problems with both the audio and video. The stuttering is synchronous in both, and occurs at seemingly random intervals. I've used PowerDVD 9, PowerDVD 11 trial, and WinDVD 11 Pro trial. All three have stuttering problems, although PowerDVD 11 seems to have the least. Watching system resource usage, CPU load is never above 20%, and memory usage tends to be a constant 1/3rd the total available system memory. When playback is fine, its superb...the video is crystal clear. The audio quality is ok, certainly not what I would expect from a BluRay disc. I did some research, and it seems that playing BluRay from a PC causes a downsampling of the audio? I am curious if the audio is my primary problem here, the cause of the stuttering I am encountering? When stuttering occurs, the audio gets REALLY bad, while the video just pauses momentarily every second until for whatever reason everything picks up and runs fine (usually after a few seconds to a couple minutes.) The audio chipset is a Realtek HD ALC887 8-channel, supposedly designed to support BluRay playback. Has anyone encountered any issues like this playing back bluray discs on a PC (namely with PowerDVD...WinDVD was FAR worse, and seemed to have real trouble even reading the discs, and I have no interest in fiddling with it further.) Is there any reason to suspect the video decoding as the problem?(Given how bad the audio gets during a stutter, and how clean the video remains, I am inclined to think the issue boils down to audio.) Is it even remotely possible that the motherboard, cpu, or ram are causing the stuttering (all three are pretty blazing fast...faster than the hardware that I replaced, which seemed to play BluRay fine with PowerDVD 9.) I've read a bit about the Asus Xonar HDAV 1.3 and the Auzen X-Fi HomeTheater HD home theater hi-fi audio cards. Seems they are the only way to get true full-quality, uncompressed BluRay audio bitstreaming over HDMI on a PC. None of the usual suspects seem to have these cards in stock, however. Are these cards worth getting? Are they even still available, or have they been discontinued (if so, that would indeed be sad...they sound simply fantastic.)

    Read the article

  • How to place SuperFetch cache on an SSD?

    - by Ian Boyd
    I'm thinking of adding a solid state drive (SSD) to my existing Windows 7 installation. I know I can (and should) move my paging file to the SSD: Should the pagefile be placed on SSDs? Yes. Most pagefile operations are small random reads or larger sequential writes, both of which are types of operations that SSDs handle well. In looking at telemetry data from thousands of traces and focusing on pagefile reads and writes, we find that Pagefile.sys reads outnumber pagefile.sys writes by about 40 to 1, Pagefile.sys read sizes are typically quite small, with 67% less than or equal to 4 KB, and 88% less than 16 KB. Pagefile.sys writes are relatively large, with 62% greater than or equal to 128 KB and 45% being exactly 1 MB in size. In fact, given typical pagefile reference patterns and the favorable performance characteristics SSDs have on those patterns, there are few files better than the pagefile to place on an SSD. What I don't know is if I even can put a SuperFetch cache (i.e. ReadyBoost cache) on the solid state drive. I want to get the benefit of Windows being able to cache gigabytes of frequently accessed data on a relativly small (e.g. 30GB) solid state drive. This is exactly what SuperFetch+ReadyBoost (or SuperFetch+ReadyDrive) was designed for. Will Windows offer (or let) me place a ReadyBoost cache on a solid state flash drive connected via SATA? A problem with the ReadyBoost cache over the ReadyDrive cache is that the ReadyBoost cache does not survive between reboots. The cache is encrypted with a per-session key, making its existing contents unusable during boot and SuperFetch pre-fetching during login. Update One I know that Windows Vista limited you to only one ReadyBoost.sfcache file (I do not know if Windows 7 removed that limitation): Q: Can use use multiple devices for EMDs? A: Nope. We've limited Vista to one ReadyBoost per machine Q: Why just one device? A: Time and quality. Since this is the first revision of the feature, we decided to focus on making the single device exceptional, without the difficulties of managing multiple caches. We like the idea, though, and it's under consideration for future versions. I also know that the 4GB limit on the cache file was a limitation of the FAT filesystem used on most USB sticks - an SSD drive would be formatted with NTFS: Q: What's the largest amount of flash that I can use for ReadyBoost? A: You can use up to 4GB of flash for ReadyBoost (which turns out to be 8GB of cache w/ the compression) Q: Why can't I use more than 4GB of flash? A: The FAT32 filesystem limits our ReadyBoost.sfcache file to 4GB Can a ReadyBoost cache on an NTFS volume be larger than 4GB? Update Two The ReadyBoost cache is encrypted with a per-boot session key. This means that the cache has to be re-built after each boot, and cannot be used to help speed boot times, or latency from login to usable. Windows ReadyDrive technology takes advantage of non-volatile (NV) memory (i.e. flash) that is incorporated with some hybrid hard drives. This flash cache can be used to help Windows boot, or resume from hibernate faster. Will Windows 7 use an internal SSD drive as a ReadyBoost/*ReadyDrive*/SuperFetch cache? Is it possible to make Windows store a SuperFetch cache (i.e. ReadyBoost) on a non-removable SSD? Is it possible to not encrypt the ReadyBoost cache, and if so will Windows 7 use the cache at boot time? See also SuperUser.com: ReadyBoost + SSD = ? Windows 7 - ReadyBoost & SSD drives? Support and Q&A for Solid-State Drives Using SDD as a cache for HDD, is there a solution? Performance increase using SSD for paging/fetch/cache or ReadyBoost? (Win7) Windows 7 To Boost SSD Performance How to Disable Nonvolatile Caching

    Read the article

  • Performance of Cluster Shared Volume file copy from SAN

    - by Sequenzia
    I am hoping someone can help me out with a strange issue. We are running a Microsoft Failover Cluster with Server 2008 R2 and an Equallogic PS4000 SAN. Our main configuration has 2 Dell Poweredge T710 Servers in the cluster. We have CSV and Quorm setup. The servers each have 10 Broadcom 1Gb NICs. Right now 4 of the NICS are on the iSCSI network for accessing the SAN. They use MPIO and the Dell HIT pack. We have 5 VMs running on each node and everything runs smooth. No noticeable performance issues or anything. From the SAN I can see the 4 iSCSI connections from each server to each volume (CSV and Quorm). Again, it seems to perform great. The problem I am running into is with backups. I have tried a few backup programs like backupchain and Veeam. The problem is both of them are very very slow to backup the VMs. For instance I have a 500GB (fixed disc) VHD that’s running on the cluster. It takes over 18 hours to backup that VHD and that’s with compression and depuping turned off which is supposed to be the fasted. We also have a separate server that is just for backups. It has a lot of directed attached storage. As part of the troubleshooting I decided to bring that server into the cluster as a node. It now has access to the CSV and can read from C:\clusterstorage\volume1 which is where our VHDs live. This backup server only has 2 NICs. 1 NIC is going to the iSCSI network and the other is just on the main network. It has Intel NICS in it without any sort of MPIO or teaming. So with the 3rd server now in the cluster I started doing some benchmarking. I have a test VHD that’s about 7GBs that’s stored in the CSV. I have tested file copying that VHD from all 3 servers to directed attached storage in the respective server. The 2 Dell servers that are the main nodes in the cluster (they house the VMs) are reading that file at about 20Mbs/Sec. Which at that rate is way to slow for the backups. The other server which only has 1 NIC to the SAN is reading at around 100Mbs/Sec. I spent a few hours on the phone with Dell today about this . We went through all kind of tests and he was pretty dumb founded. He really has no idea why that server with only 1 NIC is reading about 5 times as fast as the servers with 4 NICS and MPIO. We looked at the network utilization of the NICs while the file copy was going on. The servers with the 4 NICs had a small increase of activity during the file copy but they only went up to around 8-10% on all 4 NICs. The other server with the 1 NIC jumped up to over 80% during the file copy. I plan on doing some more testing after hours and calling Dell back tomorrow but I really am confused (and so is Dell’s support rep) why I cannot get faster file copy access to the CSV on those servers. Anyone have any input on this? Any feedback would be greatly appreciated. Thanks in advance.

    Read the article

  • How to Eliminate Tape Backup and Off-site Storage Service?

    - by Daniel Lucas
    PLEASE READ UPDATE AT THE BOTTOM. THANKS! ;) Environment Info (all Windows): 2 sites 30 servers site #1 (3TB of backup data) 5 servers site #2 (1TB of backup data) MPLS backbone tunnel connecting site #1 and site #2 Current Backup Process: Online Backup (disk-to-disk) Site #1 has a server running Symantec Backup Exec 12.5 with four 1TB USB 2.0 disks. BE jobs for full backups run nightly on all servers in site #1 to these disks. Site #2 backs up to a central file server there using software they already had when we purchased them. A BE job pulls that data nightly to site #1 and stores them on said disks. Off-site Backup (tape) Connected to our backup server is a tape drive. BE backs up the external disks to tape once a week which gets picked up by our off-site storage company. Obviously we rotate two tape libraries, one is always here and one is always there. Requirements: Eliminate the need for tape and off-site storage service by doing disk-to-disk at each site and replicating site #1 to site #2 and vice versa. Software based solution as hardware options have been too pricey (ie, SonicWall, Arkeia). Agents for Exchange, SharePoint, and SQL. Some Ideas So Far: Storage DroboPro at each site with an initial 8TB of storage (these are expandable up to 16TB at present). I like these because they are rackmountable, allow disparate drives, and have iSCSI interfaces. They are relatively cheap too. Software Symantec Backup Exec 12.5 already has all the agents and licenses we need. I'd like to keep using it unless there is a better solution, similarly priced, that does everything BE does plus deduplication and replication. Server Because there is no more need for a SCSI adapter (for tape drive) we are going to virtualize our backup server as it is currently the only physical machine save for SQL boxes. Problems: When replicating between sites we want as little data as possible to go across the pipe. There is no deduplication or compression in what I have laid out here so far. The files being replicated are BE's virtual tape libraries from our disk-to-disk backup. Because of this each of those huge files will go across the wire every week because they change every day. And Finally, the Question: Is there any software out there that does deduplication, or at least compression, to handle just our site-to-site replication? Or, looking at our setup, is there any other solution that I am missing that might be cheaper, faster, better? Thanks. Sorry so long. UPDATE 2: I've set a bounty on this question to get it more attention. I'm looking for software that will handle replication of data between two sites using the least amount of data possible (either compression, deduplication, or some other method). Something similar to rsync would work but it needs to be native to Windows and not a port involving shenanigans to get up and running. Prefer a GUI based product and I don't mind shelling out a few bones if it works. Please, answers that meet the above criteria only. If you don't think one exists or if you think I'm being to restrictive keep it to yourself. If after seven days there is no answer at all, so be it. Thanks again everyone. UPDATE 2: I really appreciate everyone coming forward with suggestions. There is no way for me to try all of these before the bounty expires. For now I'm going to let this bounty run out and whoever has the most votes will get the 100 rep points. Thanks again!

    Read the article

  • CPU Utilization LAMP stack

    - by Max
    We've got an ec2 m2.4xlarge running Magento (centos 5.6, httpd 2.2, php 5.2.17 with eaccelerator 0.9.5.3, mysql 5.1.52). Right now we're getting a large traffic spike, and our top looks like this: top - 09:41:29 up 31 days, 1:12, 1 user, load average: 120.01, 129.03, 113.23 Tasks: 1190 total, 18 running, 1172 sleeping, 0 stopped, 0 zombie Cpu(s): 97.3%us, 1.8%sy, 0.0%ni, 0.5%id, 0.0%wa, 0.0%hi, 0.0%si, 0.4%st Mem: 71687720k total, 36898928k used, 34788792k free, 49692k buffers Swap: 880737784k total, 0k used, 880737784k free, 1586524k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 2433 mysql 15 0 23.6g 4.5g 7112 S 564.7 6.6 33607:34 mysqld 24046 apache 16 0 411m 65m 28m S 26.4 0.1 0:09.05 httpd 24360 apache 15 0 410m 60m 25m S 26.4 0.1 0:03.65 httpd 24993 apache 16 0 410m 57m 21m S 26.1 0.1 0:01.41 httpd 24838 apache 16 0 428m 74m 20m S 24.8 0.1 0:02.37 httpd 24359 apache 16 0 411m 62m 26m R 22.3 0.1 0:08.12 httpd 23850 apache 15 0 411m 64m 27m S 16.8 0.1 0:14.54 httpd 25229 apache 16 0 404m 46m 17m R 10.2 0.1 0:00.71 httpd 14594 apache 15 0 404m 63m 34m S 8.4 0.1 1:10.26 httpd 24955 apache 16 0 404m 50m 21m R 8.4 0.1 0:01.66 httpd 24313 apache 16 0 399m 46m 22m R 8.1 0.1 0:02.30 httpd 25119 apache 16 0 411m 59m 23m S 6.8 0.1 0:01.45 httpd Questions: Would giving msyqld more memory help it cache queries and react faster? If so, how? Other than splitting mysql and php to separate servers (which we're about to do) is there anything else we could/should be doing? Thanks! UPDATE: Here's our my.cnf along with the output of mysqltuner. It looks like a cache problem. Thanks again! # cat /etc/my.cnf [client] port = **** socket = /var/lib/mysql/mysql.sock [mysqld] datadir=/mnt/persistent/mysql port=**** socket=/var/lib/mysql/mysql.sock key_buffer = 512M max_allowed_packet = 64M table_cache = 1024 sort_buffer_size = 8M read_buffer_size = 4M read_rnd_buffer_size = 2M myisam_sort_buffer_size = 64M thread_cache_size = 128M tmp_table_size = 128M join_buffer_size = 1M query_cache_limit = 2M query_cache_size= 64M query_cache_type = 1 max_connections = 1000 thread_stack = 128K thread_concurrency = 48 log-bin=mysql-bin server-id = 1 wait_timeout = 300 innodb_data_home_dir = /mnt/persistent/mysql/ innodb_data_file_path = ibdata1:10M:autoextend innodb_buffer_pool_size = 20G innodb_additional_mem_pool_size = 20M innodb_log_file_size = 64M innodb_log_buffer_size = 8M innodb_flush_log_at_trx_commit = 1 innodb_lock_wait_timeout = 50 innodb_thread_concurrency = 48 ft_min_word_len=3 [myisamchk] ft_min_word_len=3 key_buffer = 128M sort_buffer_size = 128M read_buffer = 2M write_buffer = 2M # ./mysqltuner.pl >> MySQLTuner 1.2.0 - Major Hayden <[email protected]> >> Bug reports, feature requests, and downloads at http://mysqltuner.com/ >> Run with '--help' for additional options and output filtering -------- General Statistics -------------------------------------------------- [--] Skipped version check for MySQLTuner script [OK] Currently running supported MySQL version 5.1.52-log [OK] Operating on 64-bit architecture -------- Storage Engine Statistics ------------------------------------------- [--] Status: +Archive -BDB +Federated +InnoDB -ISAM -NDBCluster [--] Data in MyISAM tables: 2G (Tables: 26) [--] Data in InnoDB tables: 749M (Tables: 250) [!!] Total fragmented tables: 262 -------- Security Recommendations ------------------------------------------- -------- Performance Metrics ------------------------------------------------- [--] Up for: 31d 2h 30m 38s (680M q [253.371 qps], 2M conn, TX: 4825B, RX: 236B) [--] Reads / Writes: 89% / 11% [--] Total buffers: 20.6G global + 15.1M per thread (1000 max threads) [OK] Maximum possible memory usage: 35.4G (51% of installed RAM) [OK] Slow queries: 0% (35K/680M) [OK] Highest usage of available connections: 53% (537/1000) [OK] Key buffer size / total MyISAM indexes: 512.0M/457.2M [OK] Key buffer hit rate: 100.0% (9B cached / 264K reads) [OK] Query cache efficiency: 42.3% (260M cached / 615M selects) [!!] Query cache prunes per day: 4384652 [OK] Sorts requiring temporary tables: 0% (1K temp sorts / 38M sorts) [!!] Joins performed without indexes: 100404 [OK] Temporary tables created on disk: 17% (7M on disk / 45M total) [OK] Thread cache hit rate: 99% (537 created / 2M connections) [!!] Table cache hit rate: 0% (1K open / 946K opened) [OK] Open file limit used: 9% (453/5K) [OK] Table locks acquired immediately: 99% (758M immediate / 758M locks) [OK] InnoDB data size / buffer pool: 749.3M/20.0G -------- Recommendations ----------------------------------------------------- General recommendations: Run OPTIMIZE TABLE to defragment tables for better performance Enable the slow query log to troubleshoot bad queries Adjust your join queries to always utilize indexes Increase table_cache gradually to avoid file descriptor limits Variables to adjust: query_cache_size (> 64M) join_buffer_size (> 1.0M, or always use indexes with joins) table_cache (> 1024)

    Read the article

  • Windows 7 inbuilt and 3rd party (de)fragmentation related queries

    - by Karan
    I have a pretty good idea of how files end up getting fragmented. That said, I just copied ~3,200 files of varying sizes (from a few KB to ~20GB) from an external USB HDD to an internal, freshly formatted (under Windows 7 x64), NTFS, 2TB, 5400RPM, WD, SATA, non-system (i.e. secondary) drive, filling it up 57%. Since it should have been very much possible for each file to have been stored in one contiguous block, I expected the drive to be fragmented not more than 1-2% at most after this rather lengthy exercise (unfortunately this older machine doesn't support USB 3.0). Windows 7's inbuilt defrag utility told me after a quick analysis that the drive was fragmented only 1% or so, which dovetailed neatly with my expectations. However, just out of curiosity I downloaded and ran the latest portable x64 version of Piriform's Defraggler, and was shocked to see the drive being reported as being ~85% fragmented! The portable version of Auslogics Disk Defrag also agreed with Defraggler, and both clearly expected to grind away for ~10 hours to completely defragment the drive. 1) How in blazes could the inbuilt and 3rd party defrag utils disagree so badly? I mean, 10-20% variance is probably understandable, but 1% and 85% are miles apart! This Engineering Windows 7 blog post states: In Windows XP, any file that is split into more than one piece is considered fragmented. Not so in Windows Vista if the fragments are large enough – the defragmentation algorithm was changed (from Windows XP) to ignore pieces of a file that are larger than 64MB. As a result, defrag in XP and defrag in Vista will report different amounts of fragmentation on a volume. ... [Please read the entire post so the quote is not taken out of context.] Could it simply be that the 3rd party defrag utils ignore this post-XP change and continue to use analysis algos similar to those XP used? 2) Assuming that the 3rd party utils aren't lying about the real extent of fragmentation (which Windows is downplaying post-XP), how could the files have even got fragmented so badly given they were just copied over afresh to an empty drive? 3) If vastly differing analysis algos explain the yawning gap, which do I believe? I'm no defrag fanatic for sure, but 85% is enough to make me seriously consider spending 10 hours defragging this drive. On the other hand, 1% reported by Windows' own defragger clearly implies that there is no cause for concern and defragging would actually have negative consequences (as per the post). Is Windows' assumption valid and should I just let it be, or will there be any noticeable performance gains after running one of the 3rd party utils for 10 hours straight? 4) I see that out of the box Windows 7 defrag is scheduled to run weekly. Does anyone know whether it defrags every single time, or only if its analysis reveals a fragmentation percentage over a set threshold? If the latter, what is this threshold and can it be changed, maybe via a Registry edit? Thanks for reading through (my first query on this wonderful site!) and for any helpful replies. Also, if you're answering question #3, please keep in mind that any speed increases post defragging with 3rd party utils vis-à-vis Windows' inbuilt program should not include pre-Vista (preferably pre-Win7) examples. Further, examples of programs that made your system boot faster won't help in this case, since this is a non-system drive (although one that'll still be used daily).

    Read the article

  • Parallel processing slower than sequential?

    - by zebediah49
    EDIT: For anyone who stumbles upon this in the future: Imagemagick uses a MP library. It's faster to use available cores if they're around, but if you have parallel jobs, it's unhelpful. Do one of the following: do your jobs serially (with Imagemagick in parallel mode) set MAGICK_THREAD_LIMIT=1 for your invocation of the imagemagick binary in question. By making Imagemagick use only one thread, it slows down by 20-30% in my test cases, but meant I could run one job per core without issues, for a significant net increase in performance. Original question: While converting some images using ImageMagick, I noticed a somewhat strange effect. Using xargs was significantly slower than a standard for loop. Since xargs limited to a single process should act like a for loop, I tested that, and found it to be about the same. Thus, we have this demonstration. Quad core (AMD Athalon X4, 2.6GHz) Working entirely on a tempfs (16g ram total; no swap) No other major loads Results: /media/ramdisk/img$ time for f in *.bmp; do echo $f ${f%bmp}png; done | xargs -n 2 -P 1 convert -auto-level real 0m3.784s user 0m2.240s sys 0m0.230s /media/ramdisk/img$ time for f in *.bmp; do echo $f ${f%bmp}png; done | xargs -n 2 -P 2 convert -auto-level real 0m9.097s user 0m28.020s sys 0m0.910s /media/ramdisk/img$ time for f in *.bmp; do echo $f ${f%bmp}png; done | xargs -n 2 -P 10 convert -auto-level real 0m9.844s user 0m33.200s sys 0m1.270s Can anyone think of a reason why running two instances of this program takes more than twice as long in real time, and more than ten times as long in processor time to complete the same task? After that initial hit, more processes do not seem to have as significant of an effect. I thought it might have to do with disk seeking, so I did that test entirely in ram. Could it have something to do with how Convert works, and having more than one copy at once means it cannot use processor cache as efficiently or something? EDIT: When done with 1000x 769KB files, performance is as expected. Interesting. /media/ramdisk/img$ time for f in *.bmp; do echo $f ${f%bmp}png; done | xargs -n 2 -P 1 convert -auto-level real 3m37.679s user 5m6.980s sys 0m6.340s /media/ramdisk/img$ time for f in *.bmp; do echo $f ${f%bmp}png; done | xargs -n 2 -P 1 convert -auto-level real 3m37.152s user 5m6.140s sys 0m6.530s /media/ramdisk/img$ time for f in *.bmp; do echo $f ${f%bmp}png; done | xargs -n 2 -P 2 convert -auto-level real 2m7.578s user 5m35.410s sys 0m6.050s /media/ramdisk/img$ time for f in *.bmp; do echo $f ${f%bmp}png; done | xargs -n 2 -P 4 convert -auto-level real 1m36.959s user 5m48.900s sys 0m6.350s /media/ramdisk/img$ time for f in *.bmp; do echo $f ${f%bmp}png; done | xargs -n 2 -P 10 convert -auto-level real 1m36.392s user 5m54.840s sys 0m5.650s

    Read the article

< Previous Page | 161 162 163 164 165 166 167 168 169 170 171 172  | Next Page >