none
Exchange 2013 Performance Monitor Data Collector Sets

    Question

  • I've noticed that the Exchange 2013 install created two data collector sets within Performance Monitor:

    • ExchangeDiagnosticsDailyPerformanceLog
    • ExchangeDiagnosticsPerformanceLog

    These sets are creating daily log files saved in C:\Program Files\Microsoft\Exchange Server\V15\Logging\Diagnostics\DailyPerformanceLogs that are about 500MB in size, which are filling up our system volume.

    I've tried stopping them, but they get restarted automatically. I also tried deleting them, but they get recreated automatically. Is there a way to disable these? We have a 3rd party tool that we use for performance monitoring and would like to avoid having to delete these log files on a regular basis to keep our system volume from filling up.

    As a workaround, I've changed the path where the log files get saved to a non existent drive letter, but would still like to disable the Data Collector Sets completely.

    Thanks,

    -Cory

    Friday, November 16, 2012 9:18 PM

Answers

  • Where did you stop them? Within PerFmon or within the Task Scheduler?

    If you look under Windows\PLA in the Task Scheduler Manager, you can disable the actual running tasks there. See if that works.

    • Marked as answer by Cory Wood Monday, November 19, 2012 4:44 PM
    Saturday, November 17, 2012 3:48 PM
    Moderator

All replies

  • Where did you stop them? Within PerFmon or within the Task Scheduler?

    If you look under Windows\PLA in the Task Scheduler Manager, you can disable the actual running tasks there. See if that works.

    • Marked as answer by Cory Wood Monday, November 19, 2012 4:44 PM
    Saturday, November 17, 2012 3:48 PM
    Moderator
  • Hi ,

    Maybe you can modify registry to unload them.

    More information for your reference.

    How to unload/reload performance counters on Exchange 2010:

    http://blogs.technet.com/b/mikelag/archive/2010/09/10/how-to-unload-reload-performance-counters-on-exchange-2010.aspx


    Wendy Liu

    TechNet Community Support

    Monday, November 19, 2012 9:52 AM
    Moderator
  • Andy-

    I was stopping them within PerfMon. I didn't know there were scheduled tasks for them. I disabled them in Task Scheduler and that seems to have worked.

    Thanks!

    -Cory

    Monday, November 19, 2012 4:44 PM
  • These are used by Exchange Managed Availability to monitor the health of the system. I would not disable them. You might have more of a problem if you're worrying about 500MB filling up your system disk... 

    TR

    Sunday, March 31, 2013 12:32 PM
  • If these logs are important to monitoring the health of the system, then I need a way to relocate where this logging is recorded. The system disk should be fairly static. However, this logging is adding 300-500MB of file space per day to the system disk. This logging must be relocated. Any guidance?
    Monday, April 1, 2013 1:58 PM
  • I agree with dthab. In order for me to be able to re-enable them, I'll need a way to relocate where the logs are being saved. If the logs only took up 500MB, I wouldn't think twice about expanding our drive to handle them. However, they're taking up 500MB every day.

    -Cory

    Wednesday, April 3, 2013 2:37 PM
  • We are seeing this as well.  Tony, do you know how to relocate these?  500 MB per day is not scalable.
    Monday, May 13, 2013 9:34 PM
  • As a workaround that actually seems to work, I employed the Junction utility from SysInternals.

    I stopped the Exchange Diagnostics service and the data collector sets, then created junctions to replace the default logging folders to folders on a separate volume that had more free space.  So far it's working perfectly and my C:\ drive free space has stopped dropping like a rock.

    Somehow, using Microsoft's own technology to circumvent its own shortsitedness is supremely gratifying.

    Monday, July 15, 2013 10:19 PM
  • It's not just the Diagnostic Logs, MS has enabled logging for everything (EWS, CLient Access etc etc etc) in an effort to reduce time to fix and MS techs having to wait till the next day or week to get the info required. Fair enough, but the most common cause they will find is that the drive has no space due to logging.

    We now have a 100GB C drive as standard due to this ISSUE. Needs some tweaking don't you think, so that your diagnostic info isn't the actual thing that brings down the server most. Even if you redirect them, it will mean that this stuff will eventually fill some drive. Logging without constraints means we schedule an engineer to go and delete them every Monday morning. ARE YOU SERIOUS?


    • Edited by Tony Nexus Sunday, September 8, 2013 11:40 AM
    Sunday, September 8, 2013 11:39 AM
  • We too are seeing this issue on all our multi-role exchange servers.

    500mb performance logs a day? What's up with that?

    We have found disabling the scheduled task and the performance counters is only a temporary solution, they are recreated again a few days later (Server did not restart during that period) so not sure what was causing that.

    We have now resorted to stopping Microsoft Exchange Health Manager service altogether.  

    What I'd like to know is why are these so big?  I have ran Get-EventLogLevel and the event level for each identity is mostly set to "Lowest".  


    • Edited by Andrew SP Friday, October 4, 2013 2:37 PM
    Friday, October 4, 2013 2:36 PM
  • We too are seeing this issue on all our multi-role exchange servers.

    500mb performance logs a day? What's up with that?

    We have found disabling the scheduled task and the performance counters is only a temporary solution, they are recreated again a few days later (Server did not restart during that period) so not sure what was causing that.

    We have now resorted to stopping Microsoft Exchange Health Manager service altogether.  

    What I'd like to know is why are these so big?  I have ran Get-EventLogLevel and the event level for each identity is mostly set to "Lowest".  


    It's Monday morning, I've come in and it's time to clear up some log files again.  Despite my efforts of stopping Microsoft Exchange Health Manager these scheduled tasks are still being created as are the performance counters.

    550mb logs generated per day.

    Please explain how to stop this from happening completely or to drastically decrease the size of them.


    • Edited by Andrew SP Monday, October 7, 2013 6:18 AM
    Monday, October 7, 2013 6:18 AM
  • One "quick" solution is to move the path of the log files.Please refer to following link on how to: http://www.spr33.com/exchange-2013-data-collector-sets/

    Put it on a Networkshare that wont affect other systems.

    Im having same issue and still looking for info about how long they are collected,when they Automatic purge(or if it can be adjusted).

    Will come back With info when i know how :)


    Please mark as helpful if you find my contribution useful or as an answer if it does answer your question. That will encourage me - and others - to take time out to help you. Thank you! Off2work

    • Proposed as answer by Off2work Monday, November 11, 2013 1:45 PM
    Monday, November 11, 2013 1:45 PM
  • Thanks Off2work. 

    I just wanted to make sure to express my displeasure with this issue as well.  One more disappointing thing about Exchange 2013.....

    Quick update:  I moved the diagnostics and log files, but they revert to the default after a reboot......

    Worst app I've ever dealt with, and I've dealt with some bad ones.


    Thursday, April 10, 2014 8:38 PM
  • Do note,

    These performance counters only take up 1 week and not more so it will be about 3-5 GB if you can`t spare this on you c disk you have not implemented any best practise exchange design.

    specialy with todays JBOD SATA design your c disk will be 1TB +. Review your designs to match MS advise.


    MCTS-MCITP exchange 2010 | MCTS-MCITP Exchange: 2007 | MCSA Messaging: 2003 | MCP windows 2000

    Thursday, June 5, 2014 10:21 AM
  • Do note,

    These performance counters only take up 1 week and not more so it will be about 3-5 GB if you can`t spare this on you c disk you have not implemented any best practise exchange design.

    specialy with todays JBOD SATA design your c disk will be 1TB +. Review your designs to match MS advise.


    MCTS-MCITP exchange 2010 | MCTS-MCITP Exchange: 2007 | MCSA Messaging: 2003 | MCP windows 2000

    Hello Martin,

    Many of us are implementing exchange in a virtual environment, where disk is at a premium.  Spending money on a 1 TB system drive is not a reasonable expectation.  Even if that is best practice, there should be a way to manage the location and size of these log files. 

    The inability to manage the location of log/performance data for centralization, or other reasons, is one more example of the disconnect between MS and their customer's use cases in regards to 2013. 

    These are the requirements for 2013 from MS:

    • At least 30 GB on the drive on which you install Exchange

    • An additional 500 MB of available disk space for each Unified Messaging (UM) language pack that you plan to install

    • 200 MB of available disk space on the system drive

    • A hard disk that stores the message queue database on with at least 500 MB of free space.

    So your statement is invalid as 30 GB is not even close to deal with the amount of log and performance data that we cannot modify in any way. 

    Thursday, June 5, 2014 3:37 PM
  • I hope your talking about very small deployments than, ( Less than 1000 mailboxes ). And the smalles disk i can order here is 1TB for nearline SAS so no clue what you talk about spending additional money.

    Exchange on virtual platform is more expensive than on hardware and on top of it it is not more resilliant so there is no gain in it. 

    With exchange 2013 you build small servers with cheap nearline SAS disks of 1 to 4 TB. Non of the servers needs redudant PSU or anything in it. Its not a problem anymore if a server fails.

    30 GB is more than enough all my virtual test labs run on 50 GB luns 20 Gb for the OS and 30 GB for exchange.

    Do note that it is exchange only requirements you need 30 GB on top of what you have installed allready and YES you need this for managed availibility in production environments.


    MCTS-MCITP exchange 2010 | MCTS-MCITP Exchange: 2007 | MCSA Messaging: 2003 | MCP windows 2000

    Friday, June 6, 2014 6:38 AM
  • You can use a script to clear out the old logs:

    http://www.c7solutions.com/2013/04/removing-old-exchange-2013-log-files-html

    I have plenty of space on my C:, but have still use my own script to clear logs on a weekly basis.


    Twitter!: Please Note: My Posts are provided “AS IS” without warranty of any kind, either expressed or implied.

    Friday, June 6, 2014 11:37 AM
    Moderator
  • By default Exchange will retain 5GB of those daily performance logs. That limit can be adjusted in the Micorsoft.Exchange.Diagnostics.Service.exe.config file. I would recommend retaining those files, but if you absolutely can not spare the space on your C drive, change the MaxSize=5128 attribute to a lower number.
    Friday, June 13, 2014 6:56 PM
  • Meanwhile, back in the real world, finding a 20-user Exchange server filling it's system drive with 500-1000MB/day of logging and not tidying itself up is a little bit of an issue.

    Not sure how I'm going to fix it short and long term, but I'll sort that tomorrow.

    As others say, I don't expect manufacturer diagnostic logging to (by default) take a production box down with overzealous logging to the system drive. My system drives aren't built for logging, they're built for OS's.


    Wednesday, June 25, 2014 8:51 AM
  • Meanwhile, back in the real world, finding a 20-user Exchange server filling it's system drive with 500-1000MB/day of logging and not tidying itself up is a little bit of an issue.

    Not sure how I'm going to fix it short and long term, but I'll sort that tomorrow.

    As others say, I don't expect manufacturer diagnostic logging to (by default) take a production box down with overzealous logging to the system drive. My system drives aren't built for logging, they're built for OS's.


    Its not 500/1000 MB/ Day,  

    Its only 500/1000MB/Day for the Last 5 days = max of 5 GB per server

    This means having a 50-60GB c drive is sufficiant for small deployment although i would reccoment following the Microsoft PA and user large SAS 7200 disk as OS and as database/log disk. in that case space aint a issue.


    MCTS exchange 2013 | MCTS-MCITP exchange 2010 | MCTS-MCITP Exchange: 2007 | MCSA Messaging: 2003 | MCP windows 2000

    Wednesday, June 25, 2014 9:46 AM
  • I believe that this 5 GB default limit didn't exist in RTM back when this thread was started... I can't remember for sure but I feel like these logs back then were generating far more than that...

    Regardless, good to know the value can be tweaked now.

    For the life of me I'll still never understand deploying exchange system drives without at least basic RAID1.  Database disks I can understand in high-node-volume DAGs, but what a serious pain in the ass to have to rebuild a failed server when super cheap raid1 can save the day.

    Of course, like most people here (it sounds like) we have smaller deployments (<1,000 mailboxes each) so our 4-node DAGs are completely virtualized on Hyper-V.  80 GB C drives are now doable thanks to being able to adjust this value.

    IIS logs are another one that build up rapidly - what do folks do to clear those - just script?

    Wednesday, June 25, 2014 10:58 AM