Search Results

Search found 13151 results on 527 pages for 'performance counters'.

Page 180/527 | < Previous Page | 176 177 178 179 180 181 182 183 184 185 186 187  | Next Page >

• SQL Server and Hyper-V Dynamic Memory Part 3

  - by SQLOS Team

  In parts 1 and 2 of this series we looked at the basics of Hyper-V Dynamic Memory and SQL Server memory management. In this part Serdar looks at configuration guidelines for SQL Server memory management. Part 3: Configuration Guidelines for Hyper-V Dynamic Memory and SQL Server Now that we understand SQL Server Memory Management and Hyper-V Dynamic Memory basics, let’s take a look at general configuration guidelines in order to utilize benefits of Hyper-V Dynamic Memory in your SQL Server VMs. Requirements Host Operating System Requirements Hyper-V Dynamic Memory feature is introduced with Windows Server 2008 R2 SP1. Therefore in order to use Dynamic Memory for your virtual machines, you need to have Windows Server 2008 R2 SP1 or Microsoft Hyper-V Server 2008 R2 SP1 in your Hyper-V host. Guest Operating System Requirements In addition to this Dynamic Memory is only supported in Standard, Web, Enterprise and Datacenter editions of windows running inside VMs. Make sure that your VM is running one of these editions. For additional requirements on each operating system see “Dynamic Memory Configuration Guidelines” here. SQL Server Requirements All versions of SQL Server support Hyper-V Dynamic Memory. However, only certain editions of SQL Server are aware of dynamically changing system memory. To have a truly dynamic environment for your SQL Server VMs make sure that you are running one of the SQL Server editions listed below: ·         SQL Server 2005 Enterprise ·         SQL Server 2008 Enterprise / Datacenter Editions ·         SQL Server 2008 R2 Enterprise / Datacenter Editions Configuration guidelines for other versions of SQL Server are covered below in the FAQ section. Guidelines for configuring Dynamic Memory Parameters Here is how to configure Dynamic Memory for your SQL VMs in a nutshell: Hyper-V Dynamic Memory Parameter Recommendation Startup RAM 1 GB + SQL Min Server Memory Maximum RAM > SQL Max Server Memory Memory Buffer % 5 Memory Weight Based on performance needs   Startup RAM In order to ensure that your SQL Server VMs can start correctly, ensure that Startup RAM is higher than configured SQL Min Server Memory for your VMs. Otherwise SQL Server service will need to do paging in order to start since it will not be able to see enough memory during startup. Also note that Startup Memory will always be reserved for your VMs. This will guarantee a certain level of performance for your SQL Servers, however setting this too high will limit the consolidation benefits you’ll get out of your virtualization environment. Maximum RAM This one is obvious. If you’ve configured SQL Max Server Memory for your SQL Server, make sure that Dynamic Memory Maximum RAM configuration is higher than this value. Otherwise your SQL Server will not grow to memory values higher than the value configured for Dynamic Memory. Memory Buffer % Memory buffer configuration is used to provision file cache to virtual machines in order to improve performance. Due to the fact that SQL Server is managing its own buffer pool, Memory Buffer setting should be configured to the lowest value possible, 5%. Configuring a higher memory buffer will prevent low resource notifications from Windows Memory Manager and it will prevent reclaiming memory from SQL Server VMs. Memory Weight Memory weight configuration defines the importance of memory to a VM. Configure higher values for the VMs that have higher performance requirements. VMs with higher memory weight will have more memory under high memory pressure conditions on your host. Questions and Answers Q1 – Which SQL Server memory model is best for Dynamic Memory? The best SQL Server model for Dynamic Memory is “Locked Page Memory Model”. This memory model ensures that SQL Server memory is never paged out and it’s also adaptive to dynamically changing memory in the system. This will be extremely useful when Dynamic Memory is attempting to remove memory from SQL Server VMs ensuring no SQL Server memory is paged out. You can find instructions on configuring “Locked Page Memory Model” for your SQL Servers here. Q2 – What about other SQL Server Editions, how should I configure Dynamic Memory for them? Other editions of SQL Server do not adapt to dynamically changing environments. They will determine how much memory they should allocate during startup and don’t change this value afterwards. Therefore make sure that you configure a higher startup memory for your VM because that will be all the memory that SQL Server utilize Tune Maximum Memory and Memory Buffer based on the other workloads running on the system. If there are no other workloads consider using Static Memory for these editions. Q3 – What if I have multiple SQL Server instances in a VM? Having multiple SQL Server instances in a VM is not a general recommendation for predictable performance, manageability and isolation. In order to achieve a predictable behavior make sure that you configure SQL Min Server Memory and SQL Max Server Memory for each instance in the VM. And make sure that: ·         Dynamic Memory Startup Memory is greater than the sum of SQL Min Server Memory values for the instances in the VM ·         Dynamic Memory Maximum Memory is greater than the sum of SQL Max Server Memory values for the instances in the VM Q4 – I’m using Large Page Memory Model for my SQL Server. Can I still use Dynamic Memory? The short answer is no. SQL Server does not dynamically change its memory size when configured with Large Page Memory Model. In virtualized environments Hyper-V provides large page support by default. Most of the time, Large Page Memory Model doesn’t bring any benefits to a SQL Server if it’s running in virtualized environments. Q5 – How do I monitor SQL performance when I’m trying Dynamic Memory on my VMs? Use the performance counters below to monitor memory performance for SQL Server: Process - Working Set: This counter is available in the VM via process performance counters. It represents the actual amount of physical memory being used by SQL Server process in the VM. SQL Server – Buffer Cache Hit Ratio: This counter is available in the VM via SQL Server counters. This represents the paging being done by SQL Server. A rate of 90% or higher is desirable. Conclusion These blog posts are a quick start to a story that will be developing more in the near future. We’re still continuing our testing and investigations to provide more detailed configuration guidelines with example performance numbers with a white paper in the upcoming months. Now it’s time to give SQL Server and Hyper-V Dynamic Memory a try. Use this guidelines to kick-start your environment. See what you think about it and let us know of your experiences. - Serdar Sutay Originally posted at http://blogs.msdn.com/b/sqlosteam/

  Read the article

• Issue 15: Oracle PartnerNetwork Exchange @ Oracle OpenWorld

  - by rituchhibber

       ORACLE FOCUS Oracle PartnerNetwork Exchange@ ORACLE OpenWorld Sylvie MichouSenior DirectorPartner Marketing & Communications and Strategic Programs RESOURCES -- Oracle OpenWorld 2012 Oracle PartnerNetwork Exchange @ OpenWorld Oracle PartnerNetwork Exchange @ OpenWorld Registration Oracle PartnerNetwork Exchange SpecializationTest Fest Oracle OpenWorld Schedule Builder Oracle OpenWorld Promotional Toolkit for Partners Oracle Partner Events Oracle Partner Webcasts Oracle EMEA Partner News SUBSCRIBE FEEDBACK PREVIOUS ISSUES If you are attending our forthcoming Oracle OpenWorld 2012 conference in San Francisco from 30 September to 4 October, you will discover a new dedicated programme of keynotes and sessions tailored especially for you, our valued partners. Oracle PartnerNetwork Exchange @ OpenWorld has been created to enhance the opportunities for you to learn from and network with Oracle executives and experts. The programme also provides more informal opportunities than ever throughout the week to meet up with the people who are most important to your business: customers, prospects, colleagues and the Oracle EMEA Alliances & Channels management team. Oracle remains fully focused on building the industry's most admired partner ecosystem—which today spans over 25,000 partners. This new OPN Exchange programme offers an exciting change of pace for partners throughout the conference. Now it will be possible to enjoy a fully-integrated, partner-dedicated session schedule throughout the week, as well as key social events such as the Sunday night Welcome Reception, networking lunches from Monday to Thursday at the Howard Street Tent, and a fantastic closing event on the last Thursday afternoon. In addition to the regular Oracle OpenWorld conference schedule, if you have registered for the Oracle PartnerNetwork Exchange @ OpenWorld programme, you will be invited to attend a much anticipated global partner keynote presentation, plus more than 40 conference sessions aimed squarely at what's most important to you, as partners. Prominent topics for discussion will include: Oracle technologies and roadmaps and how they fit with partners' business plans; business development; regional distinctions in business practices; and much more. Each session will provide plenty of food for thought ahead of the numerous networking opportunities throughout the week, encouraging the knowledge exchange with Oracle executives, customers, prospects, and colleagues that will make this conference of even greater value for you. At Oracle we always work closely with our partners to deliver solution offerings that improve business value, simplify the IT experience and drive innovation and efficiencies for joint customers. The most important element of our new OPN Exchange is content that helps you get more from technology investments, more from your peer-to-peer connections, and more from your interactions with customers. To this end we've created some partner-specific tools which can be used by OPN members ahead of the conference itself. Crucially, a comprehensive Content Catalog already lists and organises details of every OPN Exchange session, speaker, exhibitor, demonstration and related materials. This Content Catalog can be used by all our partners to identify interesting content that you can add to your own personalised Oracle OpenWorld Schedule Builder, allowing more effective planning and pre-enrolment for vital sessions. There are numerous highlights that you will definitely want to include in those personal schedules. On Sunday morning, 30 September we will start the week with partner dedicated OPN Exchange sessions, following our Global Partner Keynote at 13:00 with Judson Althoff, SVP, Worldwide Alliances & Channels and Embedded Sales and senior executives, giving insight into Oracle's partner vision, strategy, and resources—all designed to help build and strengthen market opportunities for you. This will be followed by a number of OPN Exchange general sessions, the Oracle OpenWorld Opening Keynote with Larry Ellison, CEO, Oracle and concluded with the OPN Exchange AfterDark Welcome Reception, starting at 19:30 at the Metreon. From Monday 1 to Thursday 4 October, you can attend the OPN Exchange sessions that are most relevant to your business today and over the coming year. Oracle's top product and sales leaders will be on hand to discuss Oracle's strategic direction in 40+ targeted and in-depth sessions focussing on critical success factors to develop your business. Oracle's dedication to innovation, specialization, enablement and engineering provides Oracle partners with a huge opportunity to create new services and solutions, differentiate themselves and deliver extreme value to joint customers across the globe. Oracle will even be helping over 1000 partners to earn OPN Specialization certification during the Oracle OpenWorld OPN Exchange Test Fest, which will be providing all the study materials and exams required to drive Specialization for free at the conference. You simply need to check the list of current certification tracks available, and make sure you pre-register to reserve a seat in one of the ten sessions being offered free to OPN Exchange registered attendees. And finally, let's not forget those all-important networking opportunities, which can so often provide partners with valuable long-term alliances as well as exciting new business leads. The Oracle PartnerNetwork Lounge, located at Moscone South, exhibition hall, room 100 is the place where partners can meet formally or informally with colleagues, customers, prospects, and other industry professionals. OPN Specialized partners with OPN Exchange passes can also visit the OPN Video Blogging room to record and share ideas, and at the OPN Information Station you will find consultants available to answer your questions. "For the first time ever we will have a full partner conference within OpenWorld. OPN Exchange @ OpenWorld will kick-off on the first Sunday and run the entire week. We'll have over 40 sessions throughout that time and partners will hear from our top development executives, with special sessions dedicated to partnering throughout. It's going to be a phenomenal event, and we look forward to seeing our partners there." Judson Althoff, SVP, Oracle Worldwide Alliances & Channels and Embedded Sales So if you haven't done so already, please register for Oracle PartnerNetwork Exchange @ OpenWorld today or add OPN Exchange to your existing registration for just $100 through My Account. And if you have any further questions regarding partner activities at Oracle OpenWorld, please don't hesitate to contact the Oracle PartnerNetwork team at [email protected] will be on hand to share the very latest information about: Oracle's SPARC Superclusters: the latest Engineered Systems from Oracle, delivering radically improved performance, faster deployment and greatly reduced operational costs for mixed database and enterprise application consolidation Oracle's SPARC T4 servers: with the newly developed T4 processor and Oracle Solaris providing up to five times the single threaded performance and better overall system throughput for expanded application versatility Oracle Database Appliance: a new way to take advantage of the world's most popular database, Oracle Database 11g, in a single, easy-to-deploy and manage system. It's a complete package engineered to deliver simple, reliable and affordable database services to small and medium size businesses and departmental systems. All hardware and software components are supported together and offer customers unique pay-as-you-grow software licensing to quickly scale from two to 24 processor cores without incurring the costs and downtime usually associated with hardware upgrades Oracle Exalogic: the world's only integrated cloud machine, featuring server hardware and middleware software engineered together for maximum performance with minimum set-up and operational cost Oracle Exadata Database Machine: the only database machine that provides extreme performance for both data warehousing and online transaction processing (OLTP) applications, making it the ideal platform for consolidating onto grids or private clouds. It is a complete package of servers, storage, networking and software that is massively scalable, secure and redundant Oracle Sun ZFS Storage Appliances: providing enterprise-class NAS performance, price-performance, manageability and TCO by combining third-generation software with high-performance controllers, flash-based caches and disks Oracle Pillar Axiom Quality-of-Service: confidently consolidate storage for multiple applications into a single datacentre storage solution Oracle Solaris 11: delivering secure enterprise cloud deployments with the ability to run hundreds of virtual application with no overhead and co-engineered with other Oracle software products to provide the highest levels of security, manageability and performance Oracle Enterprise Manager 12c: Oracle's integrated enterprise IT management product, providing the industry's only complete, integrated and business-driven enterprise cloud management solution Oracle VM 3.0: the latest release of Oracle's server virtualisation and management solution, helping to move datacentres beyond server consolidation to improve application deployment and management. Register today and ensure your place at the Extreme Performance Tour! Extreme Performance Tour events are free to attend, but places are limited. To make sure that you don't miss out, please visit Oracle's Extreme Performance Tour website, select the city that you'd be interest in attending an event in, and then click on the 'Register Now' button for that city to secure your interest. Each individual city page also contains more in-depth information about your local event, including logistics, agenda and maybe even a preview of VIP guest speakers. -- Oracle OpenWorld 2010 Whether you attended Oracle OpenWorld 2009 or not, don't forget to save the date now for Oracle OpenWorld 2010. The event will be held a little earlier next year, from 19th-23rd September, so please don't miss out. With thousands of sessions and hundreds of exhibits and demos already lined up, there's no better place to learn how to optimise your existing systems, get an inside line on upcoming technology breakthroughs, and meet with your partner peers, Oracle strategists and even the developers responsible for the products and services that help you get better results for your end customers. Register Now for Oracle OpenWorld 2010! Perhaps you are interested in learning more about Oracle OpenWorld 2010, but don't wish to register at this time? Great! Please just enter your contact information here and we will contact you at a later date. How to Exhibit at Oracle OpenWorld 2010 Sponsorship Opportunities at Oracle OpenWorld 2010 Advertising Opportunities at Oracle OpenWorld 2010 -- Back to the welcome page

  Read the article

• How can I test this SQL Server performance Utility?

  - by Martin Smith

  As part of my MSc I need to do a three month project later this year. I have decided to do something which will likely be useful for me in the workplace and spend the time getting to understand SQL Server internals. The deliverable for this project will be a performance advisor looking at a variety of different rules. Some static such as finding redundant indexes, some more dynamic such as using XEvents to find outlying invocations of stored procedure execution times when certain parameters are passed. I am struggling to come up with a good way of testing this though. I can obviously design a "bad" database and a synthetic workload that my tool will pick up issues on but I also need to demonstrate that it has real world utility. Looking at the self tuning database literature it is common to use TPC benchmarks but I've had a look at the TPCC site and it looks very time consuming to implement and not that good a fit to my project's testing needs in any event (I would still be able to "rig" it by the decisions I made on indexing or physical architecture). Plan A would be to find willing beta tester(s) but in the event that isn't possible I will need a fallback plan. The best idea I have come up with so far is to use the various MS sample applications as examples of real world applications. e.g. http://msftdpprodsamples.codeplex.com/ http://www.asp.net/community/projects/ Does anyone have any better suggestions?

  Read the article

• Why the performance of following code is degrading when I use threads ?

  - by DotNetBeginner

  Why the performance of following code is degrading when I use threads ? **1.Without threads int[] arr = new int[100000000]; //Array elements - [0][1][2][3]---[100000000-1] addWithOutThreading(arr); // Time required for this operation - 1.16 sec Definition for addWithOutThreading public void addWithOutThreading(int[] arr) { UInt64 result = 0; for (int i = 0; i < 100000000; i++) { result = result + Convert.ToUInt64(arr[i]); } Console.WriteLine("Addition = " + result.ToString()); } **2.With threads int[] arr = new int[100000000]; int part = (100000000 / 4); UInt64 res1 = 0, res2 = 0, res3 = 0, res4 = 0; ThreadStart starter1 = delegate { addWithThreading(arr, 0, part, ref res1); }; ThreadStart starter2 = delegate { addWithThreading(arr, part, part * 2, ref res2); }; ThreadStart starter3 = delegate { addWithThreading(arr, part * 2, part * 3, ref res3); }; ThreadStart starter4 = delegate { addWithThreading(arr, part * 3, part * 4, ref res4); }; Thread t1 = new Thread(starter1); Thread t2 = new Thread(starter2); Thread t3 = new Thread(starter3); Thread t4 = new Thread(starter4); t1.Start(); t2.Start(); t3.Start(); t4.Start(); t1.Join(); t2.Join(); t3.Join(); t4.Join(); Console.WriteLine("Addition = "+(res1+res2+res3+res4).ToString()); // Time required for this operation - 1.30 sec Definition for addWithThreading public void addWithThreading(int[] arr,int startIndex, int endIndex,ref UInt64 result) { for (int i = startIndex; i < endIndex; i++) { result = result + Convert.ToUInt64(arr[i]); } }

  Read the article

• How to store data in mysql, to get the fastest performance?

  - by Oden

  Hey, I'm thinking about it, witch of the following two query types would give me the fastest performance for a user messaging module inside my site: The first one i thought about is a multi table setup, witch has a connection table, and a main table. The connection table holds the connection between accounts, and the messaging table. In this case a query would look like following, to get some data of the author, and the messages he has sent: SELECT m.*, a.username FROM messages AS m LEFT JOIN connection_table ON (message_id = m.id) LEFT JOIN accounts AS a ON (account_id = a.id) WHERE m.id = '32341' Inserting into it is a little bit more "complicated". My other idea, and in my thought the better solution of this problem is that i store the data i would use in a connection table in the same table where is store the data of the mail. Sounds like i would get lots of duplicated entries, but no, because i have a field witch has text type and holds user ids like this: *24*32*249* If I want to query them, i use the mysql LIKE method. Deleting is an other problem, but for this i have one more field where i store who has deleted the post. Sad about that i don't know how to join this. So what would you recommend? Are there other ways?

  Read the article

• Application Code Redesign to reduce no. of Database Hits from Performance Perspective

  - by Rachel

  Scenario I want to parse a large CSV file and inserts data into the database, csv file has approximately 100K rows of data. Currently I am using fgetcsv to parse through the file row by row and insert data into Database and so right now I am hitting database for each line of data present in csv file so currently database hit count is 100K which is not good from performance point of view. Current Code: public function initiateInserts() { //Open Large CSV File(min 100K rows) for parsing. $this->fin = fopen($file,'r') or die('Cannot open file'); //Parsing Large CSV file to get data and initiate insertion into schema. while (($data=fgetcsv($this->fin,5000,";"))!==FALSE) { $query = "INSERT INTO dt_table (id, code, connectid, connectcode) VALUES (:id, :code, :connectid, :connectcode)"; $stmt = $this->prepare($query); // Then, for each line : bind the parameters $stmt->bindValue(':id', $data[0], PDO::PARAM_INT); $stmt->bindValue(':code', $data[1], PDO::PARAM_INT); $stmt->bindValue(':connectid', $data[2], PDO::PARAM_INT); $stmt->bindValue(':connectcode', $data[3], PDO::PARAM_INT); // Execute the statement $stmt->execute(); $this->checkForErrors($stmt); } } I am looking for a way wherein instead of hitting Database for every row of data, I can prepare the query and than hit it once and populate Database with the inserts. Any Suggestions !!! Note: This is the exact sample code that I am using but CSV file has more no. of field and not only id, code, connectid and connectcode but I wanted to make sure that I am able to explain the logic and so have used this sample code here. Thanks !!!

  Read the article

• 2-column table with two foreign keys. Performance/design question.

  - by Emanuel

  Hello everyone! I recently ran into a quite complex problem and after looking around a lot I couldn't find a solution to it. I've found answers to my questions many times before on stackoverflow.com, so I decided to post here. So I'm making a user/group managment system for a web-based project, and I'm storing all related data into a postgreSQL database. This system relies on three tables: USERS GROUPS GROUP_USERS The two first tables simply define all the users and all the groups on the site, and the last table, GROUP_USERS, stores the groups every user is part of. It only has two columns: USER_ID GROUP_ID Since every user can be a member of several groups, I decided to make a separate table for this purpose, rather than storing a comma separated column in the USERS-table. Now, both columns are foreign keys, and I want to make them both primary keys as well, this since each combination of USER_ID and GROUP_ID has to be unique, and if I give them the constraint UNIQUE pgAdmin tells me that each table should have at least one Primary key. But now I am stuck with what seems to be a lot of indexes and relations to a very small table only containing numbers. In the end, I want this table to be as fast as possible, even if containing tens of thousands of rows. Size on disk shouldn't be a problem since its just all numbers anyway, but it feels quite stupid to have a full-sized index refering to a smaller table. Should I stick with my current solution, store comma-separated values in a column in the USERS-table or is there any other solution I should be aware of. PS. I don't want to use an array-column, even if they are supported by postgreSQL. I want to be as generic as possible so I can switch database later on, if necessary. EDIT: I other words, will using a compound primary key and two foreign keys in one table with only two columns have a negative impact on performance rather than the opposite due to the size of the generated index? Thank you!

  Read the article

• What are the javascript performance tradeoffs in adding id's to dom elements?

  - by Blinky

  For example, I have a website which has a dynamically-generated segment of its dom which looks roughly like this: <div id="master"> <div class="child"> ... </div> <div class="child"> ... </div> <div class="child"> ... </div> ... <div class="child"> ... </div> </div> There might be hundreds of child elements in this manner, and they, too, might have yet more children. At present, none of them have id's except for the master. Javascript manipulates them via dom-walking. It's a little scary. I'd like to add id's, e.g.: <div id="master"> <div id="child1" class="child"> ... </div> <div id="child2" class="child"> ... </div> <div id="child3" class="child"> ... </div> ... <div id="childN" class="child"> ... </div> </div> What are the tradeoffs in adding id's to these divs? Certainly, jQuery must maintain a hash of id's, so there's going to be some sort of performance hit. Any general guidelines in when adding additional id's is a bad idea? Is jQuery so awesome that, in practice, it doesn't matter?

  Read the article

• Why does reusing arrays increase performance so significantly in c#?

  - by Willem

  In my code, I perform a large number of tasks, each requiring a large array of memory to temporarily store data. I have about 500 tasks. At the beginning of each task, I allocate memory for an array : double[] tempDoubleArray = new double[M]; M is a large number depending on the precise task, typically around 2000000. Now, I do some complex calculations to fill the array, and in the end I use the array to determine the result of this task. After that, the tempDoubleArray goes out of scope. Profiling reveals that the calls to construct the arrays are time consuming. So, I decide to try and reuse the array, by making it static and reusing it. It requires some additional juggling to figure out the minimum size of the array, requiring an extra pass through all tasks, but it works. Now, the program is much faster (from 80 sec to 22 sec for execution of all tasks). double[] tempDoubleArray = staticDoubleArray; However, I'm a bit in the dark of why precisely this works so well. Id say that in the original code, when the tempDoubleArray goes out of scope, it can be collected, so allocating a new array should not be that hard right? I ask this because understanding why it works might help me figuring out other ways to achieve the same effect, and because I would like to know in what cases allocation gives performance issues.

  Read the article

• Windows 7 extremely slow login, exchange performance, printer enumeration, etc...

  - by Jeff

  Background: I have a fresh copy of Windows 7 Professional x64 on a Dell Latitude E6500. The laptop has 8GB RAM, 250GB drive, and all Intel peripherals (net/wifi/graphics). All available Windows updates, as well as hardware drivers are installed. The IT folks where I work joined the computer to our Windows 2003-based Active Directory domain. There are no errors in any logs that we've looked at, and Group Policy templates appear to have applied properly. Problem: Every time I turn on or reboot the computer, it takes between 2 to 10 (all times are actual) minutes after successfully typing my username/password to get to my desktop. My login script does not always run. Sometimes I get a black screen, and a couple of minutes later the login script will pop up and take up to 10 minutes to complete. I can get around this by hitting cntrl-shift-esc and running explorer.exe from the Task Manager. The login script continues to hang, but I can minimize it and go on about my business. Either way, it generally throws errors prior to completing. I often get slow or failed connectivity to Exchange via Outlook. When I bring up printer dialogs, they take several minutes to populate, and block the calling app while doing so. Copies to SMB shares are very slow. On my home network, everything works fine. On both the work network and home network, I can use remote internet resources just fine. Web pages pull up, remote VPN's are fine, I can max out bandwidth on SpeakEasy Speed Test. I can get almost max bandwidth transferring FTP/HTTP over a LAN. Another symptom of the problem is that when I first log in, the work network shows as "Identifying" for a long time in the Network and Sharing Center, and will often then change to the name of the work domain, but say "Unauthenticated Network". Note that this computer previously ran Windows Vista with none of these problems. Attempts to Fix: Installed the Win7 admin pack Uninstalled/reinstalled all hardware drivers Verified Active Directory DNS settings (Vista works relatively well on the same network) Reset all TCP/IP settings on all adapters using the netsh commands to do so Disabled ipv6 on all adapters Disable wifi adapter while on work network Locked the network card to 100/Full, 1000/Full; also tried Auto Added various important addresses to hosts file (exchange, dns, ad) -- removed when didn't help My background is a jpeg (sounds unrelated but there is apparently a win7 login bug related to solid color background) More I have forgotten The IT staff at my company indicated they believe this is due to having Windows 2003 AD servers and not having any Windows 2008 R2 AD servers. Other than that, they have no advice or assistance to offer other than a rebuild (already tried that once with similar symptoms), or downgrade to Vista. Any thoughts out there?

  Read the article

• Performance of IIS+ASP.NET vs (NGINX + FastCGI + Mono or XSP)?

  - by dynback.com

  How much requests each can handle? How many RAM needed? As I remember FastCGI are opened initialized processes, each can handle one request. What about multithreading?

  Read the article

• How to allow users to monitor performance of a set of servers without touching every server?

  - by Jon Seigel

  I'm not a sysadmin, so this may be trivial. We have about 20 Windows Server 2008 R2 VMs we want to monitor centrally using Perfmon. The only issue is that the user account that's going to be doing the monitoring is not (and I assume will never be) in the Administrators group. The servers, and the user account (currently one, but could be more) are all on the same domain. Right now we're running a pilot with 5 of the servers, touching each VM manually to set the permissions, which is already getting cumbersome to manage. If we decide to roll this out to all the servers, we need a scalable solution to control access. What is the most flexible way to accomplish this? I'd like a solution that would work with 200 servers just as easily as the 20 servers we have now.

  Read the article

• Windows Server 2003 with Apache and IIS causing random faulting and performance issues with Apache?

  - by contrebis

  I'm trying to fix a problem on a Windows Server 2003 SE install which is running IIS6 and Apache webserver (with PHP and MySQL). IIS sites are bound to one IP, Apache to the other. Everything seemed fine till the other IP address was installed to allow a webservice to run under IIS. Symptoms: Apache now responds very slowly, even requests for static files (often 30 seconds or more) Sporadic errors are appearing in the event logs like: Faulting application httpd.exe, version 2.2.14.0, faulting module php5ts.dll, version 5.2.13.13, fault address 0x000ac14f. I've double-checked the config files, taken account of this question/answer http://serverfault.com/questions/51230/running-iis-and-apache-on-the-same-windows-server, upped the Apache log level to debug, run TCPView to check for conflicting bindings, upgraded to latest Apache/PHP versions but still no success or indication of a cause. Any suggestions on where to look, or debugging tips would be gratefully received. I'm a web programmer so not so familiar with Windows Server admin or details of the networking stack. Running PHP under IIS is not an option and hosting on another server is non-ideal.

  Read the article

• Improving TCP performance over a gigabit network with lots of connections and high traffic of small packets

  - by MinimeDJ

  I’m trying to improve my TCP throughput over a “gigabit network with lots of connections and high traffic of small packets”. My server OS is Ubuntu 11.10 Server 64bit. There are about 50.000 (and growing) clients connected to my server through TCP Sockets (all on the same port). 95% of of my packets have size of 1-150 bytes (TCP header and payload). The rest 5% vary from 150 up to 4096+ bytes. With the config below my server can handle traffic up to 30 Mbps (full duplex). Can you please advice best practice to tune OS for my needs? My /etc/sysctl.cong looks like this: kernel.pid_max = 1000000 net.ipv4.ip_local_port_range = 2500 65000 fs.file-max = 1000000 # net.core.netdev_max_backlog=3000 net.ipv4.tcp_sack=0 # net.core.rmem_max = 16777216 net.core.wmem_max = 16777216 net.core.somaxconn = 2048 # net.ipv4.tcp_rmem = 4096 87380 16777216 net.ipv4.tcp_wmem = 4096 65536 16777216 # net.ipv4.tcp_synack_retries = 2 net.ipv4.tcp_syncookies = 1 net.ipv4.tcp_mem = 50576 64768 98152 # net.core.wmem_default = 65536 net.core.rmem_default = 65536 net.ipv4.tcp_window_scaling=1 # net.ipv4.tcp_mem= 98304 131072 196608 # net.ipv4.tcp_timestamps = 0 net.ipv4.tcp_rfc1337 = 1 net.ipv4.ip_forward = 0 net.ipv4.tcp_congestion_control=cubic net.ipv4.tcp_tw_recycle = 0 net.ipv4.tcp_tw_reuse = 0 # net.ipv4.tcp_orphan_retries = 1 net.ipv4.tcp_fin_timeout = 25 net.ipv4.tcp_max_orphans = 8192 Here are my limits: $ ulimit -a core file size (blocks, -c) 0 data seg size (kbytes, -d) unlimited scheduling priority (-e) 0 file size (blocks, -f) unlimited pending signals (-i) 193045 max locked memory (kbytes, -l) 64 max memory size (kbytes, -m) unlimited open files (-n) 1000000 pipe size (512 bytes, -p) 8 POSIX message queues (bytes, -q) 819200 real-time priority (-r) 0 stack size (kbytes, -s) 8192 cpu time (seconds, -t) unlimited max user processes (-u) 1000000 [ADDED] My NICs are the following: $ dmesg | grep Broad [ 2.473081] Broadcom NetXtreme II 5771x 10Gigabit Ethernet Driver bnx2x 1.62.12-0 (2011/03/20) [ 2.477808] bnx2x 0000:02:00.0: eth0: Broadcom NetXtreme II BCM57711E XGb (A0) PCI-E x4 5GHz (Gen2) found at mem fb000000, IRQ 28, node addr d8:d3:85:bd:23:08 [ 2.482556] bnx2x 0000:02:00.1: eth1: Broadcom NetXtreme II BCM57711E XGb (A0) PCI-E x4 5GHz (Gen2) found at mem fa000000, IRQ 40, node addr d8:d3:85:bd:23:0c [ADDED 2] ethtool -k eth0 Offload parameters for eth0: rx-checksumming: on tx-checksumming: on scatter-gather: on tcp-segmentation-offload: on udp-fragmentation-offload: off generic-segmentation-offload: on generic-receive-offload: on large-receive-offload: on rx-vlan-offload: on tx-vlan-offload: on ntuple-filters: off receive-hashing: off [ADDED 3] sudo ethtool -S eth0|grep -vw 0 NIC statistics: [1]: rx_bytes: 17521104292 [1]: rx_ucast_packets: 118326392 [1]: tx_bytes: 35351475694 [1]: tx_ucast_packets: 191723897 [2]: rx_bytes: 16569945203 [2]: rx_ucast_packets: 114055437 [2]: tx_bytes: 36748975961 [2]: tx_ucast_packets: 194800859 [3]: rx_bytes: 16222309010 [3]: rx_ucast_packets: 109397802 [3]: tx_bytes: 36034786682 [3]: tx_ucast_packets: 198238209 [4]: rx_bytes: 14884911384 [4]: rx_ucast_packets: 104081414 [4]: rx_discards: 5828 [4]: rx_csum_offload_errors: 1 [4]: tx_bytes: 35663361789 [4]: tx_ucast_packets: 194024824 [5]: rx_bytes: 16465075461 [5]: rx_ucast_packets: 110637200 [5]: tx_bytes: 43720432434 [5]: tx_ucast_packets: 202041894 [6]: rx_bytes: 16788706505 [6]: rx_ucast_packets: 113123182 [6]: tx_bytes: 38443961940 [6]: tx_ucast_packets: 202415075 [7]: rx_bytes: 16287423304 [7]: rx_ucast_packets: 110369475 [7]: rx_csum_offload_errors: 1 [7]: tx_bytes: 35104168638 [7]: tx_ucast_packets: 184905201 [8]: rx_bytes: 12689721791 [8]: rx_ucast_packets: 87616037 [8]: rx_discards: 2638 [8]: tx_bytes: 36133395431 [8]: tx_ucast_packets: 196547264 [9]: rx_bytes: 15007548011 [9]: rx_ucast_packets: 98183525 [9]: rx_csum_offload_errors: 1 [9]: tx_bytes: 34871314517 [9]: tx_ucast_packets: 188532637 [9]: tx_mcast_packets: 12 [10]: rx_bytes: 12112044826 [10]: rx_ucast_packets: 84335465 [10]: rx_discards: 2494 [10]: tx_bytes: 36562151913 [10]: tx_ucast_packets: 195658548 [11]: rx_bytes: 12873153712 [11]: rx_ucast_packets: 89305791 [11]: rx_discards: 2990 [11]: tx_bytes: 36348541675 [11]: tx_ucast_packets: 194155226 [12]: rx_bytes: 12768100958 [12]: rx_ucast_packets: 89350917 [12]: rx_discards: 2667 [12]: tx_bytes: 35730240389 [12]: tx_ucast_packets: 192254480 [13]: rx_bytes: 14533227468 [13]: rx_ucast_packets: 98139795 [13]: tx_bytes: 35954232494 [13]: tx_ucast_packets: 194573612 [13]: tx_bcast_packets: 2 [14]: rx_bytes: 13258647069 [14]: rx_ucast_packets: 92856762 [14]: rx_discards: 3509 [14]: rx_csum_offload_errors: 1 [14]: tx_bytes: 35663586641 [14]: tx_ucast_packets: 189661305 rx_bytes: 226125043936 rx_ucast_packets: 1536428109 rx_bcast_packets: 351 rx_discards: 20126 rx_filtered_packets: 8694 rx_csum_offload_errors: 11 tx_bytes: 548442367057 tx_ucast_packets: 2915571846 tx_mcast_packets: 12 tx_bcast_packets: 2 tx_64_byte_packets: 35417154 tx_65_to_127_byte_packets: 2006984660 tx_128_to_255_byte_packets: 373733514 tx_256_to_511_byte_packets: 378121090 tx_512_to_1023_byte_packets: 77643490 tx_1024_to_1522_byte_packets: 43669214 tx_pause_frames: 228 Some info about SACK: When to turn TCP SACK off?

  Read the article

• Performance difference between MacBook Pro (2.8 GHz) vs Air (1.7 GHz)?

  - by jonathanconway

  I'm comparing these two Apple laptops: MacBook Pro (13", 2011 model): 2.8GHz dual-core Intel Core i7 processor with 4MB shared L3 cache 4GB (two 2GB SO-DIMMs) of 1333MHz DDR3 SDRAM AMD Radeon HD 6770M graphics processor with 1GB of GDDR5 memory on 2.4GHz configuration MacBook Air (13", 2011 model): 1.7GHz dual-core Intel Core i5 with 3MB shared L3 cache 4GB of 1333MHz DDR3 onboard memory Intel HD Graphics 3000 processor with 384MB of DDR3 SDRAM shared with main memory There's definitely a gap between them in terms of CPU speed and graphics, but what practical difference would this make on a day-to-day basis? On the one hand, I love the sleek, thin appearance of the Air. On the other hand, I don't want a machine that's going to be dog-slow when doing tasks such as running Virtual Machines, dual-booting to Windows and running multiple instances of Visual Studio, and maybe some light gaming. Is there going to be a major difference that makes the MacBook Pro a more attractive purchase?

  Read the article

• How to avoid bit torrent to degrade my connection performance without limiting bandwidth usage?

  - by Jader Dias

  The only way I found so far to have a normal web surfing, while the uTorrent is on, is to limit its bandwidth usage. There is a smarter way where the torrent can use the remaining bandwidth?

  Read the article

• How to avoid bit torrent to degrade my connection performance without limiting bandwidth usage?

  - by Jader Dias

  The only way I found so far to have a normal web surfing, while the uTorrent is on, is to limit its bandwidth usage. There is a smarter way where the torrent can use the remaining bandwidth?

  Read the article

• Alternatives for heapdumps creation with higher performance than jmap?

  - by Christian

  Hi, I have to create heapdumps, which works nice with jmap. My problem is, that jmap takes very long to create the heapdump file. Especially when the heap is getting bigger ( 1GB) it is taking too long. One situation as example: When the server gets into trouble with the heapspace, I want to restart it automatically and create a heapdump before the restart. This works, but takes too long to write the heapdump. This way the server is down for too long. The heapdump creation takes longer than one hour. I know about -XX:+HeapDumpOnOutOfMemoryError, but most of the time I can find the memory problem before the exception is thrown by the jvm. Is there an alternative to jmap which writes the heapdumps faster? A special solution for the example above would also be appreciated. This question is a mix between programming and system-administration, but I think I'm at the right place here.

  Read the article

• How should I monitor memory usage/performance in SunOS/Solaris?

  - by exhuma

  Last week we decided to add some SunOS (uname -a = SunOS bbs-sam-belair 5.10 Generic_127128-11 i86pc i386 i86pc) machines into our running munin instance. First off, the machines are pre-configured appliances, so, I want to avoid touching the system too much without supervision of the service provider. But adding it to munin was fairly easy by writing a small socket-service (if anyone is interested, I put it up on github: https://github.com/munin-monitoring/contrib/tree/master/tools/pypmmn) Yesterday, I implemented/adapted the required plugins for our machines. And here the questions start: First, I have not found a way to determine detailed memory usage values. I get the total memory by running prtconf | grep Memory, and the free memory using vmstat. Fiddling together a munin-plugin, gives me the following graph: This is pretty much uninformative. Compare this to the default plugin for linux nodes which has a lot more detail: Most importantly, this shows me how much memory is actually used by applications. So, first question: Is it possible to get detailed memory information on SunOS with the default system tools (i.e. not using top)? Onto the next puzzle: Seeing the graphs, I noticed activity in the "Paging in/out" graphs, even though the memory graph still has unused memory: Upon further investigation, I found out that df reports that /tmp is mounted on swap. Drilling around on the web, I understood that df will display swap, but in fact, it's mounted as a tmpfs. Now I don't know if this explains the swap activity. The default munin-plugin for solaris uses kstat -p -c misc -m cpu_stat to get these values. I find it already strange that this is using the cpu_stat module. So maybe I simply misinterpret the "paging" graphs? Second question: Do the paging graphs indicate that parts of the memory are paged to disk? Or is the activity caused by file operations in /tmp?

  Read the article

• Performance: Nginx SSL slowness or just SSL slowness in general?

  - by Mauvis Ledford

  I have an Amazon Web Services setup with an Apache instance behind Nginx with Nginx handling SSL and serving everything but the .php pages. In my ApacheBench tests I'm seeing this for my most expensive API call (which cache via Memcached): 100 concurrent calls to API call (http): 115ms (median) 260ms (max) 100 concurrent calls to API call (https): 6.1s (median) 11.9s (max) I've done a bit of research, disabled the most expensive SSL ciphers and enabled SSL caching (I know it doesn't help in this particular test.) Can you tell me why my SSL is taking so long? I've set up a massive EC2 server with 8CPUs and even applying consistent load to it only brings it up to 50% total CPU. I have 8 Nginx workers set and a bunch of Apache. Currently this whole setup is on one EC2 box but I plan to split it up and load balance it. There have been a few questions on this topic but none of those answers (disable expensive ciphers, cache ssl, seem to do anything.) Sample results below: $ ab -k -n 100 -c 100 https://URL This is ApacheBench, Version 2.3 <$Revision: 655654 $> Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/ Licensed to The Apache Software Foundation, http://www.apache.org/ Benchmarking URL.com (be patient).....done Server Software: nginx/1.0.15 Server Hostname: URL.com Server Port: 443 SSL/TLS Protocol: TLSv1/SSLv3,AES256-SHA,2048,256 Document Path: /PATH Document Length: 73142 bytes Concurrency Level: 100 Time taken for tests: 12.204 seconds Complete requests: 100 Failed requests: 0 Write errors: 0 Keep-Alive requests: 0 Total transferred: 7351097 bytes HTML transferred: 7314200 bytes Requests per second: 8.19 [#/sec] (mean) Time per request: 12203.589 [ms] (mean) Time per request: 122.036 [ms] (mean, across all concurrent requests) Transfer rate: 588.25 [Kbytes/sec] received Connection Times (ms) min mean[+/-sd] median max Connect: 65 168 64.1 162 268 Processing: 385 6096 3438.6 6199 11928 Waiting: 379 6091 3438.5 6194 11923 Total: 449 6264 3476.4 6323 12196 Percentage of the requests served within a certain time (ms) 50% 6323 66% 8244 75% 9321 80% 9919 90% 11119 95% 11720 98% 12076 99% 12196 100% 12196 (longest request)

  Read the article

• PostgresQL on Amazon EBS volume, realistic performance, or move to something more lightweight?

  - by Peck

  Hi, I'm working on a little research project, currently running as an instance on ec2, and I'm hoping to figure out whether I'm going down the right path. We, like a thousand other people, are making use of some of twitters streaming feeds to do gather some data to have fun with and my db seems to be having problems keeping up, and queries take what seems to be a very long time. I'm not a DBA by trade, so I'll just dump some info here and add more if need be. System specs: ec2 xl, 15 gigs of ram ebs: 4 100 gb drives, raid 0. The stream we're getting we're looking at around 10k inserts per minute. 3 main tables, with the users we're tracking somewhere in the neighborhood of 26M rows currently. Is this amount of inserts on this hardware too much to ask out of ebs? Should take a look at some things with less overhead like mongodb?

  Read the article

• How to best tune my SAN/Initiators for best performance?

  - by Disco

  Recent owner of a Dell PowerVault MD3600i i'm experiencing some weird results. I have a dedicated 24x 10GbE Switch (PowerConnect 8024), setup to jumbo frames 9K. The MD3600 has 2 RAID controllers, each has 2x 10GbE ethernet nics. There's nothing else on the switch; one VLAN for SAN traffic. Here's my multipath.conf defaults { udev_dir /dev polling_interval 5 selector "round-robin 0" path_grouping_policy multibus getuid_callout "/sbin/scsi_id -g -u -s /block/%n" prio_callout none path_checker readsector0 rr_min_io 100 max_fds 8192 rr_weight priorities failback immediate no_path_retry fail user_friendly_names yes # prio rdac } blacklist { device { vendor "*" product "Universal Xport" } # devnode "^sd[a-z]" } devices { device { vendor "DELL" product "MD36xxi" path_grouping_policy group_by_prio prio rdac # polling_interval 5 path_checker rdac path_selector "round-robin 0" hardware_handler "1 rdac" failback immediate features "2 pg_init_retries 50" no_path_retry 30 rr_min_io 100 prio_callout "/sbin/mpath_prio_rdac /dev/%n" } } And iscsid.conf : node.startup = automatic node.session.timeo.replacement_timeout = 15 node.conn[0].timeo.login_timeout = 15 node.conn[0].timeo.logout_timeout = 15 node.conn[0].timeo.noop_out_interval = 5 node.conn[0].timeo.noop_out_timeout = 10 node.session.iscsi.InitialR2T = No node.session.iscsi.ImmediateData = Yes node.session.iscsi.FirstBurstLength = 262144 node.session.iscsi.MaxBurstLength = 16776192 node.conn[0].iscsi.MaxRecvDataSegmentLength = 262144 After my tests; i can barely come to 200 Mb/s read/write. Should I expect more than that ? Providing it has dual 10 GbE my thoughts where to come around the 400 Mb/s. Any ideas ? Guidelines ? Troubleshooting tips ?

  Read the article

• iperf max udp multicast performance peaking at 10Mbit/s?

  - by Tom Frey

  I'm trying to test UDP multicast throughput via iperf but it seems like it's not sending more than 10Mbit/s from my dev machine: C:\> iperf -c 224.0.166.111 -u -T 1 -t 100 -i 1 -b 1000000000 ------------------------------------------------------------ Client connecting to 224.0.166.111, UDP port 5001 Sending 1470 byte datagrams Setting multicast TTL to 1 UDP buffer size: 8.00 KByte (default) ------------------------------------------------------------ [156] local 192.168.1.99 port 49693 connected with 224.0.166.111 port 5001 [ ID] Interval Transfer Bandwidth [156] 0.0- 1.0 sec 1.22 MBytes 10.2 Mbits/sec [156] 1.0- 2.0 sec 1.14 MBytes 9.57 Mbits/sec [156] 2.0- 3.0 sec 1.14 MBytes 9.55 Mbits/sec [156] 3.0- 4.0 sec 1.14 MBytes 9.56 Mbits/sec [156] 4.0- 5.0 sec 1.14 MBytes 9.56 Mbits/sec [156] 5.0- 6.0 sec 1.15 MBytes 9.62 Mbits/sec [156] 6.0- 7.0 sec 1.14 MBytes 9.53 Mbits/sec When I run it on another server, I'm getting ~80Mbit/s which is quite a bit better but still not anywhere near the 1Gbps limits that I should be getting? C:\> iperf -c 224.0.166.111 -u -T 1 -t 100 -i 1 -b 1000000000 ------------------------------------------------------------ Client connecting to 224.0.166.111, UDP port 5001 Sending 1470 byte datagrams Setting multicast TTL to 1 UDP buffer size: 8.00 KByte (default) ------------------------------------------------------------ [180] local 10.0.101.102 port 51559 connected with 224.0.166.111 port 5001 [ ID] Interval Transfer Bandwidth [180] 0.0- 1.0 sec 8.60 MBytes 72.1 Mbits/sec [180] 1.0- 2.0 sec 8.73 MBytes 73.2 Mbits/sec [180] 2.0- 3.0 sec 8.76 MBytes 73.5 Mbits/sec [180] 3.0- 4.0 sec 9.58 MBytes 80.3 Mbits/sec [180] 4.0- 5.0 sec 9.95 MBytes 83.4 Mbits/sec [180] 5.0- 6.0 sec 10.5 MBytes 87.9 Mbits/sec [180] 6.0- 7.0 sec 10.9 MBytes 91.1 Mbits/sec [180] 7.0- 8.0 sec 11.2 MBytes 94.0 Mbits/sec Anybody has any idea why this is not achieving close to link limits (1Gbps)? Thanks, Tom

  Read the article

• Improving TCP performance over a gigabit network lots of connections and high traffic for storage and streaming services

  - by Linux Guy

  I have two servers, Both servers hardware Specification are Processor : Dual Processor RAM : over 128 G.B Hard disk : SSD Hard disk Outging Traffic bandwidth : 3 Gbps network cards speed : 10 Gbps Server A : for Encoding videos Server B : for storage videos andstream videos over web interface like youtube The inbound bandwidth between two servers is 10Gbps , the outbound bandwidth internet bandwidth is 500Mpbs Both servers using public ip addresses in public and private network Both servers transfer and connection on nginx port , and the server B used for streaming media , like youtube stream videos Both servers in same network , when i do ping from Server A to Server B i got high time latency above 1.0ms , the time range time=52.7 ms to time=215.7 ms - This is the output of iftop utility 353Mb 707Mb 1.04Gb 1.38Gb 1.73Gb mqqqqqqqqqqqqqqqqqqqqqqqqqqqvqqqqqqqqqqqqqqqqqqqqqqqqqqqvqqqqqqqqqqqqqqqqqqqqqqqqqqqvqqqqqqqqqqqqqqqqqqqqqqqqqqqvqqqqqqqqqqqqqqqqqqqqqqqqqqq server.example.com => ip.address 6.36Mb 4.31Mb 1.66Mb <= 158Kb 94.8Kb 35.1Kb server.example.com => ip.address 1.23Mb 4.28Mb 1.12Mb <= 17.1Kb 83.5Kb 21.9Kb server.example.com => ip.address 395Kb 3.89Mb 1.07Mb <= 6.09Kb 109Kb 28.6Kb server.example.com => ip.address 4.55Mb 3.83Mb 1.04Mb <= 55.6Kb 45.4Kb 13.0Kb server.example.com => ip.address 649Kb 3.38Mb 1.47Mb <= 9.00Kb 38.7Kb 16.7Kb server.example.com => ip.address 5.00Mb 3.32Mb 1.80Mb <= 65.7Kb 55.1Kb 29.4Kb server.example.com => ip.address 387Kb 3.13Mb 1.06Mb <= 18.4Kb 39.9Kb 15.0Kb server.example.com => ip.address 3.27Mb 3.11Mb 1.01Mb <= 81.2Kb 64.5Kb 20.9Kb server.example.com => ip.address 1.75Mb 3.08Mb 2.72Mb <= 16.6Kb 35.6Kb 32.5Kb server.example.com => ip.address 1.75Mb 2.90Mb 2.79Mb <= 22.4Kb 32.6Kb 35.6Kb server.example.com => ip.address 3.03Mb 2.78Mb 1.82Mb <= 26.6Kb 27.4Kb 20.2Kb server.example.com => ip.address 2.26Mb 2.66Mb 1.36Mb <= 51.7Kb 49.1Kb 24.4Kb server.example.com => ip.address 586Kb 2.50Mb 1.03Mb <= 4.17Kb 26.1Kb 10.7Kb server.example.com => ip.address 2.42Mb 2.49Mb 2.44Mb <= 31.6Kb 29.7Kb 29.9Kb server.example.com => ip.address 2.41Mb 2.46Mb 2.41Mb <= 26.4Kb 24.5Kb 23.8Kb server.example.com => ip.address 2.37Mb 2.39Mb 2.40Mb <= 28.9Kb 27.0Kb 28.5Kb server.example.com => ip.address 525Kb 2.20Mb 1.05Mb <= 7.03Kb 26.0Kb 12.8Kb qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq TX: cum: 102GB peak: 1.65Gb rates: 1.46Gb 1.44Gb 1.48Gb RX: 1.31GB 24.3Mb 19.5Mb 18.9Mb 20.0Mb TOTAL: 103GB 1.67Gb 1.48Gb 1.46Gb 1.50Gb I check the transfer speed using iperf utility From Server A to Server B # iperf -c 0.0.0.2 -p 8777 ------------------------------------------------------------ Client connecting to 0.0.0.2, TCP port 8777 TCP window size: 85.3 KByte (default) ------------------------------------------------------------ [ 3] local 0.0.0.1 port 38895 connected with 0.0.0.2 port 8777 [ ID] Interval Transfer Bandwidth [ 3] 0.0-10.8 sec 528 KBytes 399 Kbits/sec My Current Connections in Server B # netstat -an|grep ":8777"|awk '/tcp/ {print $6}'|sort -nr| uniq -c 2072 TIME_WAIT 28 SYN_RECV 1 LISTEN 189 LAST_ACK 139 FIN_WAIT2 373 FIN_WAIT1 3381 ESTABLISHED 34 CLOSING Server A Network Card Information Settings for eth0: Supported ports: [ TP ] Supported link modes: 100baseT/Full 1000baseT/Full 10000baseT/Full Supported pause frame use: No Supports auto-negotiation: Yes Advertised link modes: 10000baseT/Full Advertised pause frame use: No Advertised auto-negotiation: Yes Speed: 10000Mb/s Duplex: Full Port: Twisted Pair PHYAD: 0 Transceiver: external Auto-negotiation: on MDI-X: Unknown Supports Wake-on: d Wake-on: d Current message level: 0x00000007 (7) drv probe link Link detected: yes Server B Network Card Information Settings for eth2: Supported ports: [ FIBRE ] Supported link modes: 10000baseT/Full Supported pause frame use: No Supports auto-negotiation: No Advertised link modes: 10000baseT/Full Advertised pause frame use: No Advertised auto-negotiation: No Speed: 10000Mb/s Duplex: Full Port: Direct Attach Copper PHYAD: 0 Transceiver: external Auto-negotiation: off Supports Wake-on: d Wake-on: d Current message level: 0x00000007 (7) drv probe link Link detected: yes ifconfig server A eth0 Link encap:Ethernet HWaddr 00:25:90:ED:9E:AA inet addr:0.0.0.1 Bcast:0.0.0.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:1202795665 errors:0 dropped:64334 overruns:0 frame:0 TX packets:2313161968 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:893413096188 (832.0 GiB) TX bytes:3360949570454 (3.0 TiB) lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:65536 Metric:1 RX packets:2207544 errors:0 dropped:0 overruns:0 frame:0 TX packets:2207544 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:247769175 (236.2 MiB) TX bytes:247769175 (236.2 MiB) ifconfig Server B eth2 Link encap:Ethernet HWaddr 00:25:90:82:C4:FE inet addr:0.0.0.2 Bcast:0.0.0.2 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:39973046980 errors:0 dropped:1828387600 overruns:0 frame:0 TX packets:69618752480 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:3013976063688 (2.7 TiB) TX bytes:102250230803933 (92.9 TiB) lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:65536 Metric:1 RX packets:1049495 errors:0 dropped:0 overruns:0 frame:0 TX packets:1049495 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:129012422 (123.0 MiB) TX bytes:129012422 (123.0 MiB) Netstat -i on Server B # netstat -i Kernel Interface table Iface MTU Met RX-OK RX-ERR RX-DRP RX-OVR TX-OK TX-ERR TX-DRP TX-OVR Flg eth2 9000 0 42098629968 0 2131223717 0 73698797854 0 0 0 BMRU lo 65536 0 1077908 0 0 0 1077908 0 0 0 LRU I Turn up send/receive buffers on the network card to 2048 and problem still persist I increase the MTU for server A and problem still persist and i increase the MTU for server B for better connectivity and transfer speed but it couldn't transfer at all The problem is : as you can see from iperf utility, the transfer speed from server A to server B slow when i restart network service in server B the transfer in server A at full speed, after 2 minutes , it's getting slow How could i troubleshoot slow speed issue and fix it in server B ? Notice : if there any other commands i should execute in servers for more information, so it might help resolve the problem , let me know in comments

  Read the article

• Can enabling a RAID controller's writeback cache harm overall performance?

  - by Nathan O'Sullivan

  I have an 8 drive RAID 10 setup connected to an Adaptec 5805Z, running Centos 5.5 and deadline scheduler. A basic dd read test shows 400mb/sec, and a basic dd write test shows about the same. When I run the two simultaneously, I see the read speed drop to ~5mb/sec while the write speed stays at more or less the same 400mb/sec. The output of iostat -x as you would expect, shows that very few read transactions are being executed while the disk is bombarded with writes. If i turn the controller's writeback cache off, I dont see a 50:50 split but I do see a marked improvement, somewhere around 100mb/s reads and 300mb/s writes. I've also found if I lower the nr_requests setting on the drive's queue (somewhere around 8 seems optimal) I can end up with 150mb/sec reads and 150mb/sec writes; ie. a reduction in total throughput but certainly more suitable for my workload. Is this a real phenomenon? Or is my synthetic test too simplistic? The reason this could happen seems clear enough, when the scheduler switches from reads to writes, it can run heaps of write requests because they all just land in the controllers cache but must be carried out at some point. I would guess the actual disk writes are occuring when the scheduler starts trying to perform reads again, resulting in very few read requests being executed. This seems a reasonable explanation, but it also seems like a massive drawback to using writeback cache on an system with non-trivial write loads. I've been searching for discussions around this all afternoon and found nothing. What am I missing?

  Read the article

< Previous Page | 176 177 178 179 180 181 182 183 184 185 186 187  | Next Page >