locked
Server down af ter time with high CPU RRS feed

  • Question

  • Hi,

    I have googled this problem, and it seems to be "known" as far as the problem goes but not with any solution.

    Symptoms:
    - WHS (SP2) goes down as in dead and with the disc activity light constantly on. Only power cycle is possible.
    There is no causes logged in the event logs. Apparently it dies to fast.

    - one of the systems two CPUs are constantly used (when the server is running)
    Task managers performance tab shows little activity on the left CPU and like constant max on the right CPU.
    Process list shows no processes using significant CPU ???!!! CPU is being used by a secret process, I don't like that! (Yes, I have ticked the show tasks for all users)

    - HDDs are showing symptoms of heating but still within limits.
    May be the result of WHS using them constantly.

    "Peoples in forums" are indicating that this problem is caused by WHS constantly doing some disc maintenance. Shutting down WHS storage manager reduces the right CPU to like 50%.
    I did this during diagnostics I don't like to run the server without storage manager.
    I have installed third party balancer software and that have helped the server some. Now it will not die every night (during client backup ?) but almost every night.


    A common thing I have seen with this problems seems to be "many files". A theory indicating the storage manager never finishing its household because of these many files (even when there are not many new files coming in).

    I have ran the storage conflict resolver (nothing wrong). It took many hours, there are 4 disks, 3 of these are in the server pool: 2x1Tb and 1x2Tb (the 4. disc is used for a daily backup). I had a disc crash and replaced the faulty 1Tb with a 2Tb. System disc (1Tb) is partitioned as 20Gb C: for system and the rest for data in the pool.

    Any solutions ?

    Notice there are more free space than capacity. Notice the encrypted amount used space :-(

    Good old dir /s from D:\ returns (all shares on D:, all d: disc in the server disk pool):

         Total Files Listed:
               160378 File(s) 1,790,345,816,455 bytes
               10977 Dir(s)  1,337,265,545,216 bytes free

     

    Dir indicates a total of 3Tb disk. 2x1 + 2Tb redundant would give an expected 2Tb max live data. It somehow makes sense when used is counted double as in 1Tb x 2 used leaves 2Tb free. 1Tb used + 2Tb free = 3Tb total. However whenever one uses of the "free" we will use 2x the storage for every new file. The numbers are misleading.

     

    Wednesday, November 16, 2011 1:09 AM

All replies

  • Hi,

    I have googled this problem, and it seems to be "known" as far as the problem goes but not with any solution.

    Symptoms:
    - WHS (SP2) goes down as in dead and with the disc activity light constantly on. Only power cycle is possible.
    There is no causes logged in the event logs. Apparently it dies to fast.

    - one of the systems two CPUs are constantly used (when the server is running)
    Task managers performance tab shows little activity on the left CPU and like constant max on the right CPU.
    Process list shows no processes using significant CPU ???!!! CPU is being used by a secret process, I don't like that! (Yes, I have ticked the show tasks for all users)

    - HDDs are showing symptoms of heating but still within limits.
    May be the result of WHS using them constantly.

    "Peoples in forums" are indicating that this problem is caused by WHS constantly doing some disc maintenance. Shutting down WHS storage manager reduces the right CPU to like 50%.
    I did this during diagnostics I don't like to run the server without storage manager.
    I have installed third party balancer software and that have helped the server some. Now it will not die every night (during client backup ?) but almost every night.


    A common thing I have seen with this problems seems to be "many files". A theory indicating the storage manager never finishing its household because of these many files (even when there are not many new files coming in).

    I have ran the storage conflict resolver (nothing wrong). It took many hours, there are 4 disks, 3 of these are in the server pool: 2x1Tb and 1x2Tb (the 4. disc is used for a daily backup). I had a disc crash and replaced the faulty 1Tb with a 2Tb. System disc (1Tb) is partitioned as 20Gb C: for system and the rest for data in the pool.

    Any solutions ?

    Notice there are more free space than capacity. Notice the encrypted amount used space :-(

    Good old dir /s from D:\ returns (all shares on D:, all d: disc in the server disk pool):

         Total Files Listed:
               160378 File(s) 1,790,345,816,455 bytes
               10977 Dir(s)  1,337,265,545,216 bytes free

     

    Dir indicates a total of 3Tb disk. 2x1 + 2Tb redundant would give an expected 2Tb max live data. It somehow makes sense when used is counted double as in 1Tb x 2 used leaves 2Tb free. 1Tb used + 2Tb free = 3Tb total. However whenever one uses of the "free" we will use 2x the storage for every new file. The numbers are misleading.

     



    hafardr@gmail.com
    Tuesday, November 22, 2011 3:09 AM