Search Results

Search found 13341 results on 534 pages for 'obiee performance tuning'.

Page 72/534 | < Previous Page | 68 69 70 71 72 73 74 75 76 77 78 79  | Next Page >

• Very uneven CPU utilization with SQL Server 2012 on 2 processor computer with 16 cores / processor

  - by cooplarsh

  After installing SQL Server Enterprise 2012 with the Server + Cal license model, on a computer with 2 processors each with 16 cores (and no hyperthreading involved) and putting the server under extremely heavy load the 16 cores on the first processor were very underutilized, the first 4 cores on the 2nd CPU were heavily utilized, and the last 12 cores were not used at all (because of the 20 core limit for this sql server version). Total CPU utilization was displaying as around 25%. Unfortunately, the server suffered from extremely poor performance even though if the tasks were evenly distributed across the 20 cores it wouldn't have been anywhere near as bad. The Windows Server was running on a VMWare virtual image under ESX Server, but all of the CPU was allocated to the windows server. We tried changing affinity settings (e.g., allocating most cores to CPU and the others to I/O), but that didn't help solve the performance problems. Upgrading the product edition to SQL Server Enterprise Core 2012 not only allowed the SQL Server to utilize the 12 previously unused cores on the 2nd processor, but it also resulted in a much more even distribution of tasks across all of the processors. To get through the backlog of requests cpU utilization jumped to around 90%, and then came down to around 33% once it was caught up, but performance improved dramatically since we failed over to the newly updated version And the performance issues went away. I was wondering if anyone knows what might cause SQL Server to unevenly distribute the load, relying almost exclusively on the first 4 cores of the 2nd processor that had 12 cores idle, and allocate only a few tasks to each of the 16 cores on the first processor. Also, is there any way we could have more evenly distributed the load across the 20 cores that were being used without the product edition upgrade? The flip side of that question is what did the product upgrade do that caused SQL Server to start evenly distributing the load across all of the cores that it recognized? Thanks to any insight to answer these questions and/or links that might help me better understand how to make sense of what was happenings.

  Read the article

• Why is only one Excel spreadsheet crippled, but others are fine?

  - by Dallas

  I have an inherited spreadsheet that I really don't want to rebuild at the moment. It's a simple small workbook that is small (< 200 rows that don't even reach to AA) and does nothing more than calculate some totals within the same worksheets. No macros, no external data sources, nothing beyond basic formatting of dates, numbers and strings. I see importing data from CSV/text has created many many workbook connections over time, but even if I delete them all (there were hundreds) it makes no difference in performance. Even clicking to simply change focus from cell to cell takes 10+ seconds, adorned by the spinning cursor and (Not Responding) appending to the title bar and the application locking up. The program seems to "recover" every time, but efficiency of editing this file is obviously seriously handicapped. All other files seem fine in Excel, and other programs have no apparent performance issues. I see Excel is chewing up CPU but I'm not sure how to narrow down what process or service is "clashing" with Excel. I tried the same file on other computers and performance is fine. If I turn off all start-up services and run only Excel, performance is restored... until I start using other programs and then it bogs down again. At this point, I would entertain almost any idea, theory or suggestion that helps pinpoint, solve or work around the issue.

  Read the article

• disk write cache buffer and separate power supply

  - by HugoRune

  Windows has a setting to turn off the write-cache buffer (see image) Turn off Windows write-cache buffer flushing on the device To prevent data loss, do not select this check box unless the device has a separate power supply that allows the device to flush its buffer in case of power failure. Is it feasible and economical to get such a "separate power supply" for the internal sata drives of a non-server PC? Under what name is such a power supply sold? I know that there are UPS devices that can be connected to external drives,but what is required to be able to switch this setting safely on for an internal disk? The setting has different descriptions in different version of windows Windows XP: Enable write caching on the disk This setting enables write caching in Windows to improve disk performance, but a power outage or equipment failure might result in data loss or corruption. Windows Server 2003: Enable write caching on the disk Recommended only for disks with a backup power supply. This setting further improves disk performance, but it also increases the risk of data loss if the disk loses power. Windows Vista: Enable advanced performance Recommended only for disks with a backup power supply. This setting further improves disk performance, but it also increases the risk of data loss if the disk loses power. Windows 7 and 8: Turn off Windows write-cache buffer flushing on the device To prevent data loss, do not select this check box unless the device has a separate power supply that allows the device to flush its buffer in case of power failure. This article by Raymond Chen has some more detailed information about what the setting does.

  Read the article

• SQL SERVER – Signal Wait Time Introduction with Simple Example – Wait Type – Day 2 of 28

  - by pinaldave

  In this post, let’s delve a bit more in depth regarding wait stats. The very first question: when do the wait stats occur? Here is the simple answer. When SQL Server is executing any task, and if for any reason it has to wait for resources to execute the task, this wait is recorded by SQL Server with the reason for the delay. Later on we can analyze these wait stats to understand the reason the task was delayed and maybe we can eliminate the wait for SQL Server. It is not always possible to remove the wait type 100%, but there are few suggestions that can help. Before we continue learning about wait types and wait stats, we need to understand three important milestones of the query life-cycle. Running - a query which is being executed on a CPU is called a running query. This query is responsible for CPU time. Runnable – a query which is ready to execute and waiting for its turn to run is called a runnable query. This query is responsible for Signal Wait time. (In other words, the query is ready to run but CPU is servicing another query). Suspended – a query which is waiting due to any reason (to know the reason, we are learning wait stats) to be converted to runnable is suspended query. This query is responsible for wait time. (In other words, this is the time we are trying to reduce). In simple words, query execution time is a summation of the query Executing CPU Time (Running) + Query Wait Time (Suspended) + Query Signal Wait Time (Runnable). Again, it may be possible a query goes to all these stats multiple times. Let us try to understand the whole thing with a simple analogy of a taxi and a passenger. Two friends, Tom and Danny, go to the mall together. When they leave the mall, they decide to take a taxi. Tom and Danny both stand in the line waiting for their turn to get into the taxi. This is the Signal Wait Time as they are ready to get into the taxi but the taxis are currently serving other customer and they have to wait for their turn. In other word they are in a runnable state. Now when it is their turn to get into the taxi, the taxi driver informs them he does not take credit cards and only cash is accepted. Neither Tom nor Danny have enough cash, they both cannot get into the vehicle. Tom waits outside in the queue and Danny goes to ATM to fetch the cash. During this time the taxi cannot wait, they have to let other passengers get into the taxi. As Tom and Danny both are outside in the queue, this is the Query Wait Time and they are in the suspended state. They cannot do anything till they get the cash. Once Danny gets the cash, they are both standing in the line again, creating one more Signal Wait Time. This time when their turn comes they can pay the taxi driver in cash and reach their destination. The time taken for the taxi to get from the mall to the destination is running time (CPU time) and the taxi is running. I hope this analogy is bit clear with the wait stats. You can check the Signalwait stats using following query of Glenn Berry. -- Signal Waits for instance SELECT CAST(100.0 * SUM(signal_wait_time_ms) / SUM (wait_time_ms) AS NUMERIC(20,2)) AS [%signal (cpu) waits], CAST(100.0 * SUM(wait_time_ms - signal_wait_time_ms) / SUM (wait_time_ms) AS NUMERIC(20,2)) AS [%resource waits] FROM sys.dm_os_wait_stats OPTION (RECOMPILE); Higher the Signal wait stats are not good for the system. Very high value indicates CPU pressure. In my experience, when systems are running smooth and without any glitch the Signal wait stat is lower than 20%. Again, this number can be debated (and it is from my experience and is not documented anywhere). In other words, lower is better and higher is not good for the system. In future articles we will discuss in detail the various wait types and wait stats and their resolution. Read all the post in the Wait Types and Queue series. Reference: Pinal Dave (http://blog.SQLAuthority.com) Filed under: Pinal Dave, PostADay, SQL, SQL Authority, SQL DMV, SQL Performance, SQL Query, SQL Scripts, SQL Server, SQL Tips and Tricks, SQL Wait Stats, SQL Wait Types, T SQL, Technology

  Read the article

• SQL SERVER – Single Wait Time Introduction with Simple Example – Wait Type – Day 2 of 28

  - by pinaldave

  In this post, let’s delve a bit more in depth regarding wait stats. The very first question: when do the wait stats occur? Here is the simple answer. When SQL Server is executing any task, and if for any reason it has to wait for resources to execute the task, this wait is recorded by SQL Server with the reason for the delay. Later on we can analyze these wait stats to understand the reason the task was delayed and maybe we can eliminate the wait for SQL Server. It is not always possible to remove the wait type 100%, but there are few suggestions that can help. Before we continue learning about wait types and wait stats, we need to understand three important milestones of the query life-cycle. Running - a query which is being executed on a CPU is called a running query. This query is responsible for CPU time. Runnable – a query which is ready to execute and waiting for its turn to run is called a runnable query. This query is responsible for Single Wait time. (In other words, the query is ready to run but CPU is servicing another query). Suspended – a query which is waiting due to any reason (to know the reason, we are learning wait stats) to be converted to runnable is suspended query. This query is responsible for wait time. (In other words, this is the time we are trying to reduce). In simple words, query execution time is a summation of the query Executing CPU Time (Running) + Query Wait Time (Suspended) + Query Single Wait Time (Runnable). Again, it may be possible a query goes to all these stats multiple times. Let us try to understand the whole thing with a simple analogy of a taxi and a passenger. Two friends, Tom and Danny, go to the mall together. When they leave the mall, they decide to take a taxi. Tom and Danny both stand in the line waiting for their turn to get into the taxi. This is the Signal Wait Time as they are ready to get into the taxi but the taxis are currently serving other customer and they have to wait for their turn. In other word they are in a runnable state. Now when it is their turn to get into the taxi, the taxi driver informs them he does not take credit cards and only cash is accepted. Neither Tom nor Danny have enough cash, they both cannot get into the vehicle. Tom waits outside in the queue and Danny goes to ATM to fetch the cash. During this time the taxi cannot wait, they have to let other passengers get into the taxi. As Tom and Danny both are outside in the queue, this is the Query Wait Time and they are in the suspended state. They cannot do anything till they get the cash. Once Danny gets the cash, they are both standing in the line again, creating one more Single Wait Time. This time when their turn comes they can pay the taxi driver in cash and reach their destination. The time taken for the taxi to get from the mall to the destination is running time (CPU time) and the taxi is running. I hope this analogy is bit clear with the wait stats. You can check the single wait stats using following query of Glenn Berry. -- Signal Waits for instance SELECT CAST(100.0 * SUM(signal_wait_time_ms) / SUM (wait_time_ms) AS NUMERIC(20,2)) AS [%signal (cpu) waits], CAST(100.0 * SUM(wait_time_ms - signal_wait_time_ms) / SUM (wait_time_ms) AS NUMERIC(20,2)) AS [%resource waits] FROM sys.dm_os_wait_stats OPTION (RECOMPILE); Higher the single wait stats are not good for the system. Very high value indicates CPU pressure. In my experience, when systems are running smooth and without any glitch the single wait stat is lower than 20%. Again, this number can be debated (and it is from my experience and is not documented anywhere). In other words, lower is better and higher is not good for the system. In future articles we will discuss in detail the various wait types and wait stats and their resolution. Read all the post in the Wait Types and Queue series. Reference: Pinal Dave (http://blog.SQLAuthority.com) Filed under: Pinal Dave, PostADay, SQL, SQL Authority, SQL DMV, SQL Performance, SQL Query, SQL Scripts, SQL Server, SQL Tips and Tricks, SQL Wait Stats, SQL Wait Types, T SQL, Technology

  Read the article

• Monitoring ASP.NET Application

  - by imran_ku07

      Introduction:          There are times when you may need to monitor your ASP.NET application's CPU and memory consumption, so that you can fine-tune your ASP.NET application(whether Web Form, MVC or WebMatrix). Also, sometimes you may need to see all the exceptions(and their details) of your application raising, whether they are handled or not. If you are creating an ASP.NET application in .NET Framework 4.0, then you can easily monitor your application's CPU or memory consumption and see how many exceptions your application raising. In this article I will show you how you can do this.       Description:           With .NET Framework 4.0, you can turn on the monitoring of CPU and memory consumption by setting AppDomain.MonitoringEnabled property to true. Also, in .NET Framework 4.0, you can register a callback method to AppDomain.FirstChanceException event to monitor the exceptions being thrown within your application's AppDomain. Turning on the monitoring and registering a callback method will add some additional overhead to your application, which will hurt your application performance. So it is better to turn on these features only if you have following properties in web.config file,   <add key="AppDomainMonitoringEnabled" value="true"/> <add key="FirstChanceExceptionMonitoringEnabled" value="true"/>             In case if you wonder what does FirstChanceException mean. It simply means the first notification of an exception raised by your application. Even CLR invokes this notification before the catch block that handles the exception. Now just update global.asax.cs file as,   string _item = "__RequestExceptionKey"; protected void Application_Start() { SetupMonitoring(); } private void SetupMonitoring() { bool appDomainMonitoringEnabled, firstChanceExceptionMonitoringEnabled; bool.TryParse(ConfigurationManager.AppSettings["AppDomainMonitoringEnabled"], out appDomainMonitoringEnabled); bool.TryParse(ConfigurationManager.AppSettings["FirstChanceExceptionMonitoringEnabled"], out firstChanceExceptionMonitoringEnabled); if (appDomainMonitoringEnabled) { AppDomain.MonitoringIsEnabled = true; } if (firstChanceExceptionMonitoringEnabled) { AppDomain.CurrentDomain.FirstChanceException += (object source, FirstChanceExceptionEventArgs e) => { if (HttpContext.Current == null)// If no context available, ignore it return; if (HttpContext.Current.Items[_item] == null) HttpContext.Current.Items[_item] = new RequestException { Exceptions = new List<Exception>() }; (HttpContext.Current.Items[_item] as RequestException).Exceptions.Add(e.Exception); }; } } protected void Application_EndRequest() { if (Context.Items[_item] != null) { //Only add the request if atleast one exception is raised var reqExc = Context.Items[_item] as RequestException; reqExc.Url = Request.Url.AbsoluteUri; Application.Lock(); if (Application["AllExc"] == null) Application["AllExc"] = new List<RequestException>(); (Application["AllExc"] as List<RequestException>).Add(reqExc); Application.UnLock(); } }               Now browse to Monitoring.cshtml file, you will see the following screen,                            The above screen shows you the total bytes allocated, total bytes in use and CPU usage of your application. The above screen also shows you all the exceptions raised by your application which is very helpful for you. I have uploaded a sample project on github at here. You can find Monitoring.cshtml file on this sample project. You can use this approach in ASP.NET MVC, ASP.NET WebForm and WebMatrix application.       Summary:          This is very important for administrators/developers to manage and administer their web application after deploying to production server. This article will help administrators/developers to see the memory and CPU usage of their web application. This will also help administrators/developers to see all the exceptions your application is throwing whether they are swallowed or not. Hopefully you will enjoy this article too.   SyntaxHighlighter.all()

  Read the article

• Using Event Driven Programming in games, when is it beneficial?

  - by Arthur Wulf White

  I am learning ActionScript 3 and I see the Event flow adheres to the W3C recommendations. From what I learned events can only be captured by the dispatcher unless, the listener capturing the event is a DisplayObject on stage and a parent of the object firing the event. You can capture the events in the capture(before) or bubbling(after) phase depending on Listner and Event setup you use. Does this system lend itself well for game programming? When is this system useful? Could you give an example of a case where using events is a lot better than going without them? Are they somehow better for performance in games? Please do not mention events you must use to get a game running, like Event.ENTER_FRAME Or events that are required to get input from the user like, KeyboardEvent.KEY_DOWN and MouseEvent.CLICK. I am asking if there is any use in firing events that have nothing to do with user input, frame rendering and the likes(that are necessary). I am referring to cases where objects are communicating. Is this used to avoid storing a collection of objects that are on the stage? Thanks Here is some code I wrote as an example of event behavior in ActionScript 3, enjoy. package regression { import flash.display.Shape; import flash.display.Sprite; import flash.events.Event; import flash.events.EventDispatcher; import flash.events.KeyboardEvent; import flash.events.MouseEvent; import flash.events.EventPhase; /** * ... * @author ... */ public class Check_event_listening_1 extends Sprite { public const EVENT_DANCE : String = "dance"; public const EVENT_PLAY : String = "play"; public const EVENT_YELL : String = "yell"; private var baby : Shape = new Shape(); private var mom : Sprite = new Sprite(); private var stranger : EventDispatcher = new EventDispatcher(); public function Check_event_listening_1() { if (stage) init(); else addEventListener(Event.ADDED_TO_STAGE, init); } private function init(e:Event = null):void { trace("test begun"); addChild(mom); mom.addChild(baby); stage.addEventListener(EVENT_YELL, onEvent); this.addEventListener(EVENT_YELL, onEvent); mom.addEventListener(EVENT_YELL, onEvent); baby.addEventListener(EVENT_YELL, onEvent); stranger.addEventListener(EVENT_YELL, onEvent); trace("\nTest1 - Stranger yells with no bubbling"); stranger.dispatchEvent(new Event(EVENT_YELL, false)); trace("\nTest2 - Stranger yells with bubbling"); stranger.dispatchEvent(new Event(EVENT_YELL, true)); stage.addEventListener(EVENT_PLAY, onEvent); this.addEventListener(EVENT_PLAY, onEvent); mom.addEventListener(EVENT_PLAY, onEvent); baby.addEventListener(EVENT_PLAY, onEvent); stranger.addEventListener(EVENT_PLAY, onEvent); trace("\nTest3 - baby plays with no bubbling"); baby.dispatchEvent(new Event(EVENT_PLAY, false)); trace("\nTest4 - baby plays with bubbling"); baby.dispatchEvent(new Event(EVENT_PLAY, true)); trace("\nTest5 - baby plays with bubbling but is not a child of mom"); mom.removeChild(baby); baby.dispatchEvent(new Event(EVENT_PLAY, true)); mom.addChild(baby); stage.addEventListener(EVENT_DANCE, onEvent, true); this.addEventListener(EVENT_DANCE, onEvent, true); mom.addEventListener(EVENT_DANCE, onEvent, true); baby.addEventListener(EVENT_DANCE, onEvent); trace("\nTest6 - Mom dances without bubbling - everyone is listening during capture phase(not target and bubble phase)"); mom.dispatchEvent(new Event(EVENT_DANCE, false)); trace("\nTest7 - Mom dances with bubbling - everyone is listening during capture phase(not target and bubble phase)"); mom.dispatchEvent(new Event(EVENT_DANCE, true)); } private function onEvent(e : Event):void { trace("Event was captured"); trace("\nTYPE : ", e.type, "\nTARGET : ", objToName(e.target), "\nCURRENT TARGET : ", objToName(e.currentTarget), "\nPHASE : ", phaseToString(e.eventPhase)); } private function phaseToString(phase : int):String { switch(phase) { case EventPhase.AT_TARGET : return "TARGET"; case EventPhase.BUBBLING_PHASE : return "BUBBLING"; case EventPhase.CAPTURING_PHASE : return "CAPTURE"; default: return "UNKNOWN"; } } private function objToName(obj : Object):String { if (obj == stage) return "STAGE"; else if (obj == this) return "MAIN"; else if (obj == mom) return "Mom"; else if (obj == baby) return "Baby"; else if (obj == stranger) return "Stranger"; else return "Unknown" } } } /*result : test begun Test1 - Stranger yells with no bubbling Event was captured TYPE : yell TARGET : Stranger CURRENT TARGET : Stranger PHASE : TARGET Test2 - Stranger yells with bubbling Event was captured TYPE : yell TARGET : Stranger CURRENT TARGET : Stranger PHASE : TARGET Test3 - baby plays with no bubbling Event was captured TYPE : play TARGET : Baby CURRENT TARGET : Baby PHASE : TARGET Test4 - baby plays with bubbling Event was captured TYPE : play TARGET : Baby CURRENT TARGET : Baby PHASE : TARGET Event was captured TYPE : play TARGET : Baby CURRENT TARGET : Mom PHASE : BUBBLING Event was captured TYPE : play TARGET : Baby CURRENT TARGET : MAIN PHASE : BUBBLING Event was captured TYPE : play TARGET : Baby CURRENT TARGET : STAGE PHASE : BUBBLING Test5 - baby plays with bubbling but is not a child of mom Event was captured TYPE : play TARGET : Baby CURRENT TARGET : Baby PHASE : TARGET Test6 - Mom dances without bubbling - everyone is listening during capture phase(not target and bubble phase) Event was captured TYPE : dance TARGET : Mom CURRENT TARGET : STAGE PHASE : CAPTURE Event was captured TYPE : dance TARGET : Mom CURRENT TARGET : MAIN PHASE : CAPTURE Test7 - Mom dances with bubbling - everyone is listening during capture phase(not target and bubble phase) Event was captured TYPE : dance TARGET : Mom CURRENT TARGET : STAGE PHASE : CAPTURE Event was captured TYPE : dance TARGET : Mom CURRENT TARGET : MAIN PHASE : CAPTURE */

  Read the article

• SQL SERVER – Backing Up and Recovering the Tail End of a Transaction Log – Notes from the Field #042

  - by Pinal Dave

  [Notes from Pinal]: The biggest challenge which people face is not taking backup, but the biggest challenge is to restore a backup successfully. I have seen so many different examples where users have failed to restore their database because they made some mistake while they take backup and were not aware of the same. Tail Log backup was such an issue in earlier version of SQL Server but in the latest version of SQL Server, Microsoft team has fixed the confusion with additional information on the backup and restore screen itself. Now they have additional information, there are a few more people confused as they have no clue about this. Previously they did not find this as a issue and now they are finding tail log as a new learning. Linchpin People are database coaches and wellness experts for a data driven world. In this 42nd episode of the Notes from the Fields series database expert Tim Radney (partner at Linchpin People) explains in a very simple words, Backing Up and Recovering the Tail End of a Transaction Log. Many times when restoring a database over an existing database SQL Server will warn you about needing to make a tail end of the log backup. This might be your reminder that you have to choose to overwrite the database or could be your reminder that you are about to write over and lose any transactions since the last transaction log backup. You might be asking yourself “What is the tail end of the transaction log”. The tail end of the transaction log is simply any committed transactions that have occurred since the last transaction log backup. This is a very crucial part of a recovery strategy if you are lucky enough to be able to capture this part of the log. Most organizations have chosen to accept some amount of data loss. You might be shaking your head at this statement however if your organization is taking transaction logs backup every 15 minutes, then your potential risk of data loss is up to 15 minutes. Depending on the extent of the issue causing you to have to perform a restore, you may or may not have access to the transaction log (LDF) to be able to back up those vital transactions. For example, if the storage array or disk that holds your transaction log file becomes corrupt or damaged then you wouldn’t be able to recover the tail end of the log. If you do have access to the physical log file then you can still back up the tail end of the log. In 2013 I presented a session at the PASS Summit called “The Ultimate Tail Log Backup and Restore” and have been invited back this year to present it again. During this session I demonstrate how you can back up the tail end of the log even after the data file becomes corrupt. In my demonstration I set my database offline and then delete the data file (MDF). The database can’t become more corrupt than that. I attempt to bring the database back online to change the state to RECOVERY PENDING and then backup the tail end of the log. I can do this by specifying WITH NO_TRUNCATE. Using NO_TRUNCATE is equivalent to specifying both COPY_ONLY and CONTINUE_AFTER_ERROR. It as its name says, does not try to truncate the log. This is a great demo however how could I achieve backing up the tail end of the log if the failure destroys my entire instance of SQL and all I had was the LDF file? During my demonstration I also demonstrate that I can attach the log file to a database on another instance and then back up the tail end of the log. If I am performing proper backups then my most recent full, differential and log files should be on a server other than the one that crashed. I am able to achieve this task by creating new database with the same name as the failed database. I then set the database offline, delete my data file and overwrite the log with my good log file. I attempt to bring the database back online and then backup the log with NO_TRUNCATE just like in the first example. I encourage each of you to view my blog post and watch the video demonstration on how to perform these tasks. I really hope that none of you ever have to perform this in production, however it is a really good idea to know how to do this just in case. It really isn’t a matter of “IF” you will have to perform a restore of a production system but more of a “WHEN”. Being able to recover the tail end of the log in these sever cases could be the difference of having to notify all your business customers of data loss or not. If you want me to take a look at your server and its settings, or if your server is facing any issue we can Fix Your SQL Server. Note: Tim has also written an excellent book on SQL Backup and Recovery, a must have for everyone. Reference: Pinal Dave (http://blog.sqlauthority.com)Filed under: Notes from the Field, PostADay, SQL, SQL Authority, SQL Performance, SQL Query, SQL Server, SQL Tips and Tricks, T SQL

  Read the article

• SQL SERVER – #TechEdIn – Presenting Tomorrow on Speed Up! – Parallel Processes and Unparalleled Performance at TechEd India 2012

  - by pinaldave

  Performance tuning is always a very hot topic when it is about SQL Server. SQL Server Performance Tuning is a very challenging subject that requires expertise in Database Administration and Database Development. I always have enjoyed talking about SQL Server Performance tuning subject. However, in India, it’s actually the very first time someone is presenting on this interesting subject, so this time I had the biggest challenge to present this session. Frequently enough, we get these two kind of questions: How to turn off parallelism as it is reducing performance? How to turn on parallelism as I want more performance? The reality is that not everyone knows what exactly is needed by their system. In this session, I have attempted to answer this very question. I’ve decided to provide a balanced view but stay away from theory, which leads us to say “It depends”. The session will have a clear message about this towards its end. Deck Details Slides: 45+ Demos: 7+ Bonus Quiz: 5 Images: 10+ Session delivery time: 52 Mins + 8 Mins of Q & A I have presented this session a couple of times to my friends and so far have received good feedback. Oftentimes, when people hear that I am going to present 45 slides, they all say it is too much to cover. However, when I am done with the session the usual reaction is that I truly gave justice to those slides. Action Item Here are a few of the action items for all of those who are going to attend this session: If you want to attend the session, just come early. There’s a good chance that you may not get a seat because right before me, there is a session from SQL Guru Vinod Kumar. He performs a powerful delivery of million concepts in just a little time. Quiz. I will be asking few questions during the session as well as before the session starts. If you get the correct answer, I will give unique learning material for you. You may not want to miss this learning opportunity at any cosst. Session Details Title: Speed Up! – Parallel Processes and Unparalleled Performance (Add to Calendar) Abstract: “More CPU, More Performance” – A  very common understanding is that usage of multiple CPUs can improve the performance of the query. To get a maximum performance out of any query, one has to master various aspects of the parallel processes. In this deep-dive session, we will explore this complex subject with a very simple interactive demo. Attendees will walk away with proper understanding of CX_PACKET wait types, MAXDOP, parallelism threshold and various other concepts. Date and Time: March 23, 2012, 12:15 to 13:15 Location: Hotel Lalit Ashok - Kumara Krupa High Grounds, Bengaluru – 560001, Karnataka, India. Add to Calendar Reference: Pinal Dave (http://blog.sqlauthority.com) Filed under: PostADay, SQL, SQL Authority, SQL Interview Questions and Answers, SQL Query, SQL Server, SQL Tips and Tricks, SQLServer, T SQL, Technology

  Read the article

• Skype performance in IPSEC VPN

  - by dunxd

  I've been challenged to "improve Skype performance" for calls within my organisation. Having read the Skype IT Administrators Guide I am wondering whether we might have a performance issue where the Skype Clients in a call are all on our WAN. The call is initiated by a Skype Client at our head office, and terminated on a Skype Client in a remote office connected via IPSEC VPN. Where this happens, I assume the trafficfrom Client A (encrypted by Skype) goes to our ASA 5510, where it is furtehr encrypted, sent to the remote ASA 5505 decrypted, then passed to Client B which decrypts the Skype encryption. Would the call quality benefit if the traffic didn't go over the VPN, but instead only relied on Skype's encryption? I imagine I could achieve this by setting up a SOCKS5 proxy in our HQ DMZ for Skype traffic. Then the traffic goes from Client A to Proxy, over the Skype relay network, then arrives at Cisco ASA 5505 as any other internet traffic, and then to Client B. Is there likely to be any performance benefit in doing this? If so, is there a way to do it that doesn't require a proxy? Has anyone else tackled this?

  Read the article

• How do I improve my incremental-backup performance?

  - by Alistair Bell

  I'm currently using the traditional rsync+cp -al method to create incremental/snapshot backups of our server tree. The backups are going onto a pair of eight-disk towers connected to the backup machine (a Sandy Bridge machine with 16 GB of RAM, running CentOS 5.5) via four eSATA connections (four disks per connection). Each disk is a regular 2 TB disk, so we have 32 TB of disk space connected to the backup machine. We're backing up about 20 TB of data on the servers with this. The problem is that each daily backup is taking more than 24 hours, and the real time-killer isn't the actual rsync, but the time it takes to perform a cp -al of the tree locally on the backup machine. It's taking more than 12 hours just to make the shadow copy of the tree, and as far as I can tell the performance backlog is at the disk (top shows the cp using a lot of RAM but not a lot of CPU and mostly in uninterruptible-sleep state) We have the server data split into four major volumes (and a few minor ones), and each of these backups runs in parallel (with some offsets in the cron to try to get some disks' cp done first). There are two volumes on the backup drive, both striped LVM volumes of 16 TB each. So obviously I need to improve the performance because it's unusable as it stands. The first question is: when CentOS 6 comes out, with support for btrfs, will making snapshots of subvolumes with btrfs substantially increase this performance? The second is: is there a way, with ext3 or something else supported in CentOS 5 or 6, to 'encourage' it to put the directories/inodes in one part of a volume (which could happen to be the part that's on an SSD, via LVM) and the files in another? That would presumably solve the problem, but I don't know of ways to hint ext3 like that.

  Read the article

• Poor SSL performance with vsftpd

  - by petrus

  I'm trying to tweak vsftpd to achieve maximum performance for my usage: I have only one or two clients that connect to the server. File size is between ~15MB and 1GB. Typical transfer batch represent between 1 and 2GB of data. For testing purposes, I'm using a tmpfs on both sides (thus eliminating any disks bottleneck) with a single 1GB file. When SSL is disabled, performance is good, with a transfer rate at ~120MB/s (reaching the limits of gigabit networking). With SSL enabled only for control traffic (and not data traffic), performance drops at about 112MB/s, which is still within the acceptable limits. However, when SSL is enabled for data flows, the transfer speed drops dramatically: 6.7MB/s using 3DES & SHA (ssl_ciphers=DES-CBC3-SHA in vsftpd.conf) 16MB/s using DES & SHA (ssl_ciphers=DES-CBC-SHA) I didn't tested other ciphers, but from what I can see from the CPU usage during the transfer, it seems that vsftpd is only using a single cpu/core per client. While this can fit for large ftp sites with hundreds of clients, I'd like to avoid this behavior and use more ressources on the server. On a side note, if you have any ideas regarding other openssl ciphers...

  Read the article

• Performance degrades for more than 2 threads on Xeon X5355

  - by zoolii

  Hi All, I am writing an application using boost threads and using boost barriers to synchronize the threads. I have two machines to test the application. Machine 1 is a core2 duo (T8300) cpu machine (windows XP professional - 4GB RAM) where I am getting following performance figures : Number of threads :1 , TPS :21 Number of threads :2 , TPS :35 (66 % improvement) further increase in number of threads decreases the TPS but that is understandable as the machine has only two cores. Machine 2 is a 2 quad core ( Xeon X5355) cpu machine (windows 2003 server with 4GB RAM) and has 8 effective cores. Number of threads :1 , TPS :21 Number of threads :2 , TPS :27 (28 % improvement) Number of threads :4 , TPS :25 Number of threads :8 , TPS :24 As you can see, performance is degrading after 2 threads (though it has 8 cores). If the program has some bottle neck , then for 2 thread also it should have degraded. Any idea? , Explanations ? , Does the OS has some role in performance ? - It seems like the Core2duo (2.4GHz) scales better than Xeon X5355 (2.66GHz) though it has better clock speed. Thank you -Zoolii

  Read the article

• Improve wireless performance

  - by djechelon

  Hello, I have a Trust Speedshare Turbo Pro router, which is running on channel 6. I found that the wireless signal (and network performance) dramatically drops from my PDA (I can barely attach to the network, even if I set the PDA's energy settings to maximum wireless performance) when I even exit my room, and I don't have shielded walls or something like that. I can't even stream a SD video from my desktop (connected via LAN) to my laptop using WiFi, while via LAN it works fine. I read that changing router's channel could improve performance due to interference reducing. I found that almost all wireless networks around here run on channels 6 and 11. I tried to go to my router's settings page to change channel, but I found that the combo box only allows me to select 6!! I'm not sure, but I may have been able in the past to change channel, though not to all of the available channels. A few minutes ago I tried a firmware upgrade, but it didn't solve my problem. My question is Is it possible that my router is someway locked to its channel? I bought it on my own, I didn't receive it from my ISP Apart from boosting the antenna power to the maximum (which, by the way, increases the EM radiation my and my family's bodies absorb 24/7 and is little more environment-unfriendly), do you have any tips on getting high quality transmission up to 5 metres from the antenna? Thank you

  Read the article

• Developing high-performance and scalable zend framework website [on hold]

  - by Daniel

  We are going to develop an ads website like http://www.gumtree.com/ (it will not be like this one but just to give you an ideea) and we are having some issues regarding performance and scalability. We are planning on using Zend Framework for this project but this is all that I'm sure off at this point. I don't think a classic approch like Zend Framework (PHP) + MySQL + Memcache + jQuery (and I would throw Doctrine 2 in there to) will fix result in a high-performance application. I was thinking on making this a RESTful application (with Zend Framework) + NGINX (or maybe MongoDB) + Memcache (or eAccelerator -- I understand this will create problems with scalability on multiple servers) + jQuery or maybe throw Backbone.js in there, a CDN for static content, a server for images and a scalable server for the requests and the rest. My questions are: - What do you think about my approch? - What solutions would you recommand for developing an high performance, scalable application expected to have a lot of traffic using PHP(Zend Framework 2)...I would be interested in your approch. I should note that I'm a Zend developer, I'm working with Zend for over 3 years, this is why I'm choosing it.

  Read the article

• WSUS performance for unneeded updates

  - by mhouston100

  We have a WSUS server serving around 300 PC's and a couple of dozen servers and a discussion came up at work as to what products to include. We have a single SQL 2005 instance on one of the servers and it has NEVER been updated. My first thought was to just tick the box for SQL 2005 and let WSUS do it's thing to upgrade to the latest service pack at least. One of the other guys here has the opinion that having updates that are relevant to only a small selection of hosts would effect the performance of WSUS as a whole, claiming that each update does a 'check' against all the hosts or something similar. My argument is that manually updating these servers is obviously not working as the admins are not paying attention to what is needed. So my question is: Do updates that only effect a sub-set of the hosts effect the overall performance of the WSUS server in relation to ALL the hosts? (disk space is not an issue at this point) Is there any performance justification for or against manually updating small amounts of products? Basically I'm needing a rebuttal against his argument and I'm unable to find any concrete documentation to prove him wrong.

  Read the article

• Hyperthreading vs. SQL Server & PostgreSQL

  - by IanC

  I have read that hyperthreading is a "performance killer" when it comes to DBs. However, what I read didn't state which CPUs. Further, it mostly indicated that I/O was "cut to < 10% performance". That logically doesn't make sense since I/O is primarily a function of controllers and disks, not CPUs. But then no one ever said bugs made sense. What I read also stated that SQL Server could put two parallel query ops onto 1 logical core (2 threads), thereby degrading performance. I have a hard time believing SQL Server's architects would have made such an obvious miscalculation. Does anyone have and data on how hyperthreading on current generation CPUs affects either of the RDBMSs I mentioned?

  Read the article

• aligning truecrypt partition on 1.5TB 4kB sector drive

  - by pQd

  hi, aligning partitions to start at real physical sector of ssds / stripped raids / 4kB drives is a 'good thing to do'. but i've run into a problems when trying to do it for a truecrypt partition that will contain ext3 on it. or so it seems. when drive is question is partitioned properly and formatted with ext3 i get very reasonable write speeds around 70-80MB/s, but when i put truecrypt and ext3 on the top of it write performance becomes very unstable and goes between 1-25MB/s with very high io-wait. on the same server i dont have any performance issues with ext3 on the top of truecrypt on regular 512B-sector 500GB sata disks. so my best guess is that iowaits are caused by misalignment but i cannot really find reliable information on how to calculate optimal partition beginning. i've tried to start it at 128 logical sector, i've also tried 8132 sector as suggested here but both gave me very bad and unstable performance. do you have any experience with similar setup? thanks!

  Read the article

• What is Peformance Monitor telling me when my page faults / second are high?

  - by David Robison

  I have a Windows 7 x64 computer that is having performance issues. After some investigation, I have discovered that the page faults / second on it, as reported by Performance Monitor, are really high. Everything else seems to be normal. Resource Monitor reports no hard faults and lots of available memory. Is this a potential cause for problems, or is it a red herring? If it is something that could be causing problems, what should I do next to figure out what is causing it? Here is a screen shot of the Performance Monitor. Notice that the average page faults / second is 75,887. On another computer that does not have problems, this number is closer to 3,000. Here is a screen shot of the Resource Monitor, sorted by hard faults / second, which is currently 0 for all processes.

  Read the article

• When to use Nginx PHP Fast CGI with a TCP socket instead of a UNIX socket?

  - by user64204

  I've followed this guide to setup PHP in FastCGI mode with Nginx. This guide describes 2 ways of doing it: TCP socket and UNIX socket. I've ran some Apache Benchmark on my locale machine and here are the results: Below tests ran multiple times to get better average statistics: $ ab -c 200 -n 100000 http://.... APACHE: 1800 req/sec NGINX (TCP socket): 2500 req/sec NGINX (UNIX socket): 15000 req/sec As far as I understand, there is overhead with using a TCP socket rather than a UNIX socket, hence the better performance with the latter. However I was not expecting such a performance difference given that the TCP socket is on the localhost, and therefore would like to ask the following question: Q: Given the huge performance gain with using a UNIX socket, what are the configuration scenarios where it would make sense to use a TCP socket instead?

  Read the article

• Xen Disk Performence Issues

  - by user98651

  I'm currently using Xen PV on CentOS 5 with my domU's as flat files running on a hardware RAID controlled (write cache enabled) formatted with XFS. On the dom0 I can get about 500MB/s in a 2GB dd write from /dev/zero however on the domU's I'm lucky if I get 10MB/s (it is usually around half that). I've tried changing the disk scheduling to NOOP on the domU's, changed some mount parameters and tweaked the performance allocations of both the dom0 (prioritize CPU) and domU's (increase RAM and VCPU allocations). None of these steps have produced any noticeable change in performance. My instinct here is that it is not a hardware problem, due to the solid performance of the dom0. Any ideas on what might be causing this problem? I'm considering moving to LVM based domU's.

  Read the article

• Processor upgrade on a laptop vs. Ram upgrade. Also does ram always matter?

  - by Evan

  I have a Dell Inspiron 14r (N4110) with an Intel Core i3 and 4gb of ram. It runs very smoothly, however gaming on this laptop is very limited. This is mostly because of integrated graphics but i have seen a computer with a Core i5, and very similar specs otherwise, run games that the N4110 cannot. This other computer has integrated graphics and 6gb of ram. I am wondering whether upgrading ram or upgrading the processor make the most difference in performance. Which setup would get better performance, an i5 with 4gb of ram or an i3 with 8 gb of ram? (Both with integrated graphics) Also, is there a certain point at which you have too much ram for the computer to ever possibly use? For instance is there really any difference in performance between 8 gb and 16 gb of ram?

  Read the article

• Perfmon quick rundown

  - by anon

  I've known quite along while about performance monitor on Windows. I have decided now to create a scheduled performance monitoring of my entire system so I can find bottlenecks for future improvements. So as you can imagine this is going to run 24/7 so I can identify peak utilization. With performance monitor on Windows 7 for example where are the logs stored (c:\perfmon)? Is there a log size? Even better is the a website that can get me up to speed with scheduling and best practices of perfmon? (I don't need an explanation of what I can monitor)

  Read the article

• SQL SERVER – Concurrency Basics – Guest Post by Vinod Kumar

  - by pinaldave

  This guest post is by Vinod Kumar. Vinod Kumar has worked with SQL Server extensively since joining the industry over a decade ago. Working on various versions from SQL Server 7.0, Oracle 7.3 and other database technologies – he now works with the Microsoft Technology Center (MTC) as a Technology Architect. Let us read the blog post in Vinod’s own voice. Learning is always fun when it comes to SQL Server and learning the basics again can be more fun. I did write about Transaction Logs and recovery over my blogs and the concept of simplifying the basics is a challenge. In the real world we always see checks and queues for a process – say railway reservation, banks, customer supports etc there is a process of line and queue to facilitate everyone. Shorter the queue higher is the efficiency of system (a.k.a higher is the concurrency). Every database does implement this using checks like locking, blocking mechanisms and they implement the standards in a way to facilitate higher concurrency. In this post, let us talk about the topic of Concurrency and what are the various aspects that one needs to know about concurrency inside SQL Server. Let us learn the concepts as one-liners: Concurrency can be defined as the ability of multiple processes to access or change shared data at the same time. The greater the number of concurrent user processes that can be active without interfering with each other, the greater the concurrency of the database system. Concurrency is reduced when a process that is changing data prevents other processes from reading that data or when a process that is reading data prevents other processes from changing that data. Concurrency is also affected when multiple processes are attempting to change the same data simultaneously. Two approaches to managing concurrent data access: Optimistic Concurrency Model Pessimistic Concurrency Model Concurrency Models Pessimistic Concurrency Default behavior: acquire locks to block access to data that another process is using. Assumes that enough data modification operations are in the system that any given read operation is likely affected by a data modification made by another user (assumes conflicts will occur). Avoids conflicts by acquiring a lock on data being read so no other processes can modify that data. Also acquires locks on data being modified so no other processes can access the data for either reading or modifying. Readers block writer, writers block readers and writers. Optimistic Concurrency Assumes that there are sufficiently few conflicting data modification operations in the system that any single transaction is unlikely to modify data that another transaction is modifying. Default behavior of optimistic concurrency is to use row versioning to allow data readers to see the state of the data before the modification occurs. Older versions of the data are saved so a process reading data can see the data as it was when the process started reading and not affected by any changes being made to that data. Processes modifying the data is unaffected by processes reading the data because the reader is accessing a saved version of the data rows. Readers do not block writers and writers do not block readers, but, writers can and will block writers. Transaction Processing A transaction is the basic unit of work in SQL Server. Transaction consists of SQL commands that read and update the database but the update is not considered final until a COMMIT command is issued (at least for an explicit transaction: marked with a BEGIN TRAN and the end is marked by a COMMIT TRAN or ROLLBACK TRAN). Transactions must exhibit all the ACID properties of a transaction. ACID Properties Transaction processing must guarantee the consistency and recoverability of SQL Server databases. Ensures all transactions are performed as a single unit of work regardless of hardware or system failure. A – Atomicity C – Consistency I – Isolation D- Durability Atomicity: Each transaction is treated as all or nothing – it either commits or aborts. Consistency: ensures that a transaction won’t allow the system to arrive at an incorrect logical state – the data must always be logically correct.  Consistency is honored even in the event of a system failure. Isolation: separates concurrent transactions from the updates of other incomplete transactions. SQL Server accomplishes isolation among transactions by locking data or creating row versions. Durability: After a transaction commits, the durability property ensures that the effects of the transaction persist even if a system failure occurs. If a system failure occurs while a transaction is in progress, the transaction is completely undone, leaving no partial effects on data. Transaction Dependencies In addition to supporting all four ACID properties, a transaction might exhibit few other behaviors (known as dependency problems or consistency problems). Lost Updates: Occur when two processes read the same data and both manipulate the data, changing its value and then both try to update the original data to the new value. The second process might overwrite the first update completely. Dirty Reads: Occurs when a process reads uncommitted data. If one process has changed data but not yet committed the change, another process reading the data will read it in an inconsistent state. Non-repeatable Reads: A read is non-repeatable if a process might get different values when reading the same data in two reads within the same transaction. This can happen when another process changes the data in between the reads that the first process is doing. Phantoms: Occurs when membership in a set changes. It occurs if two SELECT operations using the same predicate in the same transaction return a different number of rows. Isolation Levels SQL Server supports 5 isolation levels that control the behavior of read operations. Read Uncommitted All behaviors except for lost updates are possible. Implemented by allowing the read operations to not take any locks, and because of this, it won’t be blocked by conflicting locks acquired by other processes. The process can read data that another process has modified but not yet committed. When using the read uncommitted isolation level and scanning an entire table, SQL Server can decide to do an allocation order scan (in page-number order) instead of a logical order scan (following page pointers). If another process doing concurrent operations changes data and move rows to a new location in the table, the allocation order scan can end up reading the same row twice. Also can happen if you have read a row before it is updated and then an update moves the row to a higher page number than your scan encounters later. Performing an allocation order scan under Read Uncommitted can cause you to miss a row completely – can happen when a row on a high page number that hasn’t been read yet is updated and moved to a lower page number that has already been read. Read Committed Two varieties of read committed isolation: optimistic and pessimistic (default). Ensures that a read never reads data that another application hasn’t committed. If another transaction is updating data and has exclusive locks on data, your transaction will have to wait for the locks to be released. Your transaction must put share locks on data that are visited, which means that data might be unavailable for others to use. A share lock doesn’t prevent others from reading but prevents them from updating. Read committed (snapshot) ensures that an operation never reads uncommitted data, but not by forcing other processes to wait. SQL Server generates a version of the changed row with its previous committed values. Data being changed is still locked but other processes can see the previous versions of the data as it was before the update operation began. Repeatable Read This is a Pessimistic isolation level. Ensures that if a transaction revisits data or a query is reissued the data doesn’t change. That is, issuing the same query twice within a transaction cannot pickup any changes to data values made by another user’s transaction because no changes can be made by other transactions. However, this does allow phantom rows to appear. Preventing non-repeatable read is a desirable safeguard but cost is that all shared locks in a transaction must be held until the completion of the transaction. Snapshot Snapshot Isolation (SI) is an optimistic isolation level. Allows for processes to read older versions of committed data if the current version is locked. Difference between snapshot and read committed has to do with how old the older versions have to be. It’s possible to have two transactions executing simultaneously that give us a result that is not possible in any serial execution. Serializable This is the strongest of the pessimistic isolation level. Adds to repeatable read isolation level by ensuring that if a query is reissued rows were not added in the interim, i.e, phantoms do not appear. Preventing phantoms is another desirable safeguard, but cost of this extra safeguard is similar to that of repeatable read – all shared locks in a transaction must be held until the transaction completes. In addition serializable isolation level requires that you lock data that has been read but also data that doesn’t exist. Ex: if a SELECT returned no rows, you want it to return no. rows when the query is reissued. This is implemented in SQL Server by a special kind of lock called the key-range lock. Key-range locks require that there be an index on the column that defines the range of values. If there is no index on the column, serializable isolation requires a table lock. Gets its name from the fact that running multiple serializable transactions at the same time is equivalent of running them one at a time. Now that we understand the basics of what concurrency is, the subsequent blog posts will try to bring out the basics around locking, blocking, deadlocks because they are the fundamental blocks that make concurrency possible. Now if you are with me – let us continue learning for SQL Server Locking Basics. Reference: Pinal Dave (http://blog.sqlauthority.com) Filed under: PostADay, SQL, SQL Authority, SQL Performance, SQL Query, SQL Server, SQL Tips and Tricks, T SQL, Technology Tagged: Concurrency

  Read the article

• Demantra 7.3.1.3 Controlling MDP_MATRIX Combinations Assigned to Forecasting Tasks Using TargetTaskSize

  - by user702295

  New 7.3.1.3 parameter: TargetTaskSize Old parameter: BranchID  Multiple, deprecated  7.3.1.3 onwards Parameter Location: Parameters > System Parameters > Engine > Proport   Default: 0   Engine Mode: Both   Details: Specifies how many MDP_MATRIX combinations the analytical engine attempts to assign to each forecasting task.  Allocation will be affected by forecsat tree branch size.  TaskTargetSize is automcatically calculated.  It holds the perferred branch size, in number of combinations in the lowest level. This parameter is adjusted to a lower value for smaller schemas, depending on the number of available engines.   - As the forecast is generated the engine goes up the tree using max_fore_level and not top_level -1.  Max_fore_level has     to be less than or equal to top_level -1.  Due to this requirement, combinations falling under the same top level -1     member must be in the same task.  A member of the top level -1 of the forecast tree is known as a branch.  An engine     task is therefore comprised of one or more branches.     - Reveal current task size       go to Engine Administrator --> View --> Branch Information and run the application on your Demantra schema.  This will be deprecated in 7.3.1.3 since there is no longer a means of adjusting the brach size directly.  The focus is now on proper hierarchy / forecast design.     - Control of tasks       The number of tasks created is the lowest of number of branches, as defined by top level -1 members in forecast       tree, and engine sessions and the value of TargetTaskSize.  You are used to using the branch multiplier in this       calculation.  As of 7.3.1.3, the branch ID multiple is deprecated.     - Discovery of current branch size       To resolve this you must review the 2nd highest level in the forecast tree (below highest/highest) as this is the       level which determines the size of the branches.  If a few resulting tasks are too large it is recommended that       the forecast tree level driving branches be revised or at times completely removed from the forecast tree.     - Control of foreacast tree branch size         - Run the following sql to determine how even the branches are being split by the engine:             select count(*),branch_id from mdp_matrix where prediction_status = 1 and do_fore = 1 group by branch_id;             This will give you an understanding if some of the individual branches have an unusually large number of           rows and thus might indicate that the engine is not efficiently dividing up the parallel tasks.         - Based on the results of this sql, we may want to adjust the branch id multiplier and/or the number of engines           (both of these settings are found in the Engine Administrator)           select count(*), level_id from mdp_matrix where prediction_status = 1 and do_fore = 1 group by level_id;           This will give us an understanding at which level of the Forecast tree where the forecast is being generated.            Having a majority of combinations higher on the forecast tree might indicate either a poorly designed forecast           tree and/or engine parameters that are too strict           Based on the results of this we would adjust the Forecast Tree to see if choosing a different hierarchy might           produce a forecast, with more combinations, at a lower level.           For example:             - Review the 2nd highest level in the forecast tree, below highest/highest, as this is the level which               determines the size of the branches.             - If a few resulting tasks are too large it is recommended that the forecast tree level driving branches               be revised or at times completely removed from the forecast tree.               - For example, if the highest level of the forecast tree is set to Brand/All Locations.             - You have 10 brands but 2 of the brands account for 67% and 29% of all combinations.             - There is a distinct possibility that the tasks resulting from these 2 branches will be too large for               a single engine to process.  Some possible solutions could be to remove the Brand level and instead               use a different product grouping which has a more even distribution, possibly Product Group.               - It is also possible to add a location dimension to this forecast tree level, for example Customer.                This will also reduce forecast tree branch size and will deliver a balanced task allocation.             - A correctly configured Forecast Tree is something that is done by the Implementation team and is               not the responsibility of Oracle Support.  Allocation will be affected by forecast tree branch size.  When TargetTaskSize is set to 0, the default value, the system automatically calculates a value for 'TargetTaskSize' depending on the number of engines.   - QUESTION:  Does this mean that if TargetTaskSize is 1, we use tree branch size to allocate branches to tasks instead                of automatically calculating the size?     ANSWER: DEV Strongly recommends that the setting of TargetTaskSize remain at the DEFAULT of ZERO (0).   - How to control the number of engines?     Determine how many CPUs are on the machine(s) that is (are) running the engine.  As mentioned earlier, the general     rule is that you should designate 2 engines per each CPU that is available.  So for example, if you are running the     engine on a machine that has 4 CPU then you can have up to 8 engines designated in the Engine Administrator.  In this     type of architecture then instead of having one 'localhost' in your Engine Settings Screen, you would have 'localhost'     repeated eight times in this field.     Where do I set the number of engines?                 To add multiples computers where engine will run, please do a back-up of Settings.xml file under         Analytical Engines\bin\ folder, then edit it and add there the selected machines.                 Example, this will allow 3 engines to start:         - <Entry>           <Key argument="ComputerNames" />           <Value type="string" argument="localhost,localhost,localhost" />           </Entry Otherwise, if there are no additional engines defined, the calculated value of 'TargetTaskSize' is used. (Oracle does not recommend changing the default value.) The TargetTaskSize holds the engines prefered branch size, in number of level 1 combinations.   - Level 1 combinations, known as group size The engine manager will use this parameter to attempt creating branches with similar size.   * The engine manager will not create engines that do not have a branch. The engine divider algorithm uses the value of 'TargetTaskSize' as a system-preferred branch size to create branches that are more equal in size which improves engine performance.  The engine divider will try to add as many tasks as possible to an existing branch, up to the limit of 'TargetTaskSize' level 1 combinations, before adding new branches. Coming up next: - The engine divider - Group size - Level 1 combinations - MAX_FORE_LEVEL - Engine Parameters  

  Read the article

< Previous Page | 68 69 70 71 72 73 74 75 76 77 78 79  | Next Page >