locked
Should We defrag Servers ? RRS feed

  • Question

  • Hi Everybody,

    As you have lots of experience in server technology so i need a your help. I have a few questions

    1) What things i should know about Disk Drive related Servers /

    2) should we defrag the servers ?

    3) Is there  server or roles where we shouldn't run defrag.

    4) Is there any bad effect on server of defrag ?

    5) what is smart disk failure ? what we can do to resolve smart disk failure ?

    6) How we can monitor Disk problems ? like events id etc.

    I know there is few artical and post are avialable regarding smart disk failures but i didn't get . I think experience is more powerful. i need to your help.

    Thanks


    Saturday, December 8, 2018 5:49 PM

Answers

  • We debated the defrag question at my last place of employment. At the time we came to the conclusion that "it all depends". For a "classic server", that is a physical server that had it's own drive or drive array, that defrag was ok.

    Our VM's though were backed by HP 3PAR storage system which had multiple tiers of device storage. Tier 1 was SSD, tier 2 were high speed drives, and  tier 3 were slower higher capacity drives.

    We noticed that "defrag /?" now had new switches. (They may have there for a while, we just never noticed.) 

           /D      Perform traditional defrag (this is the default).  On a tiered volume
                   though, traditional defrag is performed only on the Capacity tier.
           /G      Optimize the storage tiers on the specified volumes.
           /K      Perform slab consolidation on the specified volumes.
           /L      Perform retrim on the specified volumes.
           /O      Perform the proper optimization for each media type.

    We could not figure out if our VM's could "see" the tiers or not. In the end we chose to let the storage team and the 3PAR device manage itself not do a server defrag. One concern was that a defrag would pull data from tier 3 onto tier 1 and essentially invalidate its caching mechanism.  

    https://blogs.technet.microsoft.com/askcore/2014/02/17/whats-new-in-defrag-for-windows-server-20122012r2/

    So I guess the answer is: it all depends on the back end storage, how badly fragmented the drive is, and how well the server is performing I/O wise.  

    SMART is just hardware built into the drives for health monitoring. https://en.wikipedia.org/wiki/S.M.A.R.T. 

    Monitoring the drives depends on your hardware. For HP, use the Insight Manager, for Dell use OpenManage. There should be settings in there to send email or text alerts for hardware events. 

     

    Sunday, December 9, 2018 4:59 PM
  • Hi Barinder,

    Thanks for posting here.

     1. It depends on the server, the software running on it and whether you're having any performance issues not attributed to hardware/software issues.

    We can only justify the need for defrag if it is causing some serious performance issues, else it is not worth the cost. 

    we can use tool contig.exe to check the fragmentation level before we decide to defrag.

    For your reference: 


    What's New in Defrag for Windows Server 2012/2012R2

    2. DONT defrag your Exchange Server's drive that has the databases on it, check for each vendor for each type of server before you defrag

    3. If we can back up, i do not think there is bad effect on this.

    4. A S.M.A.R.T. error is a prediction that the diagnostic test will soon fail. so we should back up as soon as possible if we have this warning. we need to learn to tell what behind this information, please check this:

    https://www.backblaze.com/blog/what-smart-stats-indicate-hard-drive-failures/

    5. we can use disk integrity checking to monitor all hard disk access to determine if the disk accurately stores information.

    Best regards,


    Please remember to mark the replies as answers if they help.
    If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com.

    Monday, December 10, 2018 8:54 AM

All replies

  • Some info here.

    https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/defrag

     

     



    Regards, Dave Patrick ....
    Microsoft Certified Professional
    Microsoft MVP [Windows Server] Datacenter Management

    Disclaimer: This posting is provided "AS IS" with no warranties or guarantees, and confers no rights.

    Saturday, December 8, 2018 6:09 PM
  • We debated the defrag question at my last place of employment. At the time we came to the conclusion that "it all depends". For a "classic server", that is a physical server that had it's own drive or drive array, that defrag was ok.

    Our VM's though were backed by HP 3PAR storage system which had multiple tiers of device storage. Tier 1 was SSD, tier 2 were high speed drives, and  tier 3 were slower higher capacity drives.

    We noticed that "defrag /?" now had new switches. (They may have there for a while, we just never noticed.) 

           /D      Perform traditional defrag (this is the default).  On a tiered volume
                   though, traditional defrag is performed only on the Capacity tier.
           /G      Optimize the storage tiers on the specified volumes.
           /K      Perform slab consolidation on the specified volumes.
           /L      Perform retrim on the specified volumes.
           /O      Perform the proper optimization for each media type.

    We could not figure out if our VM's could "see" the tiers or not. In the end we chose to let the storage team and the 3PAR device manage itself not do a server defrag. One concern was that a defrag would pull data from tier 3 onto tier 1 and essentially invalidate its caching mechanism.  

    https://blogs.technet.microsoft.com/askcore/2014/02/17/whats-new-in-defrag-for-windows-server-20122012r2/

    So I guess the answer is: it all depends on the back end storage, how badly fragmented the drive is, and how well the server is performing I/O wise.  

    SMART is just hardware built into the drives for health monitoring. https://en.wikipedia.org/wiki/S.M.A.R.T. 

    Monitoring the drives depends on your hardware. For HP, use the Insight Manager, for Dell use OpenManage. There should be settings in there to send email or text alerts for hardware events. 

     

    Sunday, December 9, 2018 4:59 PM
  • Hi Barinder,

    Thanks for posting here.

     1. It depends on the server, the software running on it and whether you're having any performance issues not attributed to hardware/software issues.

    We can only justify the need for defrag if it is causing some serious performance issues, else it is not worth the cost. 

    we can use tool contig.exe to check the fragmentation level before we decide to defrag.

    For your reference: 


    What's New in Defrag for Windows Server 2012/2012R2

    2. DONT defrag your Exchange Server's drive that has the databases on it, check for each vendor for each type of server before you defrag

    3. If we can back up, i do not think there is bad effect on this.

    4. A S.M.A.R.T. error is a prediction that the diagnostic test will soon fail. so we should back up as soon as possible if we have this warning. we need to learn to tell what behind this information, please check this:

    https://www.backblaze.com/blog/what-smart-stats-indicate-hard-drive-failures/

    5. we can use disk integrity checking to monitor all hard disk access to determine if the disk accurately stores information.

    Best regards,


    Please remember to mark the replies as answers if they help.
    If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com.

    Monday, December 10, 2018 8:54 AM
  • Thanks for your valuable help.  I will keep in mind these things.
    Saturday, December 15, 2018 7:37 PM
  • Thanks motox80. 
    Saturday, December 15, 2018 7:38 PM
  • Hi Barinder,

    Glad my advice can help you.

    Best regards,


    Please remember to mark the replies as answers if they help.
    If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com.

    Friday, December 21, 2018 5:48 AM