Informix

Informix

Connect with Db2, Informix, Netezza, open source, and other data experts to gain value from your data, share insights, and solve problems.

 View Only
  • 1.  SHMTOTAL maxed on Workgroup Edition

    Posted Tue April 18, 2023 03:07 PM

    Hi,

    I am running 14.10.FC7W1WE on Linux 64-bit VM.

    The SHMTOTAL has reached the license restriction of 32G and now Onbar Level 0 backups will not run because onbar cannot grab enough memory.

    13:20:26  Warning: The IBM IDS Workgroup Edition license restriction limits
    13:20:26  the total shared memory size for this server to 33554432 KB.
    13:20:26  Size of resident + virtual segments 413696 KB + 34189312 KB > 33554432 KB.
    13:20:26  Size of resident + virtual segments 413696KB + 34189312KB > 33554432KB
    13:20:26          total allowed by configuration parameter SHMTOTAL
    13:20:26  CDR 625001 byte memory allocation failed (Pool:arc_800992 Thread:ontape).
    13:20:26  Assert Warning: ISAM error:  An error has occurred during archive back up.

    Memory segments are as follows:

    Segment Summary:
    id         key        addr             size             ovhd     class blkused  blkfree 
    65539      525d4801   44000000         423624704        5405944  R*    103310   114     
    65540      525d4802   5d400000         2147483648       25167624 V*    459065   65223   
    65541      525d4803   dd400000         6610223104       1        B*    1613824  0       
    65542      525d4804   267400000        2692743168       1        B*    657408   0       
    65543      525d4805   307c00000        2671771648       1        B*    652288   0       
    65544      525d4806   3a7000000        12926844928      1        B*    3155968  0       
    65545      525d4807   6a9800000        561152           7848     M     137      0       
    65546      525d4808   6a9889000        561152           7848     M     135      2       
    65547      525d4809   6a9a00000        201326592        1        B*    49152    0       
    65548      525d480a   6b5a00000        201326592        1        B*    49152    0       
    65549      525d480b   6c1a00000        67108864         1        B*    16384    0       
    65550      525d480c   6c5a00000        201326592        1        B*    49152    0       
    65551      525d480d   6d1a00000        67108864         1        B*    16384    0       
    65552      525d480e   6d5a00000        402653184        1        B*    98304    0       
    65553      525d480f   6eda00000        402653184        1        B*    98304    0       
    65554      525d4810   705a00000        402653184        1        B*    98304    0       
    65555      525d4811   71da00000        402653184        1        B*    98304    0       
    65556      525d4812   735a00000        402653184        1        B*    98304    0       
    65557      525d4813   74da00000        402653184        1        B*    98304    0       
    65558      525d4814   765a00000        402653184        1        B*    98304    0       
    65559      525d4815   77da00000        805306368        1        B*    196608   0       
    65560      525d4816   7ada00000        805306368        1        B*    196608   0       
    65561      525d4817   7dda00000        805306368        1        B*    196608   0       
    65562      525d4818   80da00000        805306368        1        B*    196608   0       
    65563      525d4819   83da00000        67108864         1        B*    16384    0       
    65564      525d481a   841a00000        40820736         1        B*    9966     0       
    Total:     -          -                34359738368      -        -     8323269  65339

    I have tried onmode -F but this will not release Bufferpool segments.

    I have tried onmode -B reset and while all buffer cache is cleared it still does not allow onbar backup to run.

    IBM Support has only suggested to reduce the buffer pool settings and restart the instance.

    Is there a way to reduce the buffer pool settings to reduce the SHMTOTAL so that a full backup can run?

    Buffer pools are set as follows:

    BUFFERPOOL size=2K,start_memory=6gb,memory=8gb
    BUFFERPOOL size=4K,start_memory=2500mb,memory=3gb
    BUFFERPOOL size=8K,start_memory=2500mb,memory=3gb
    BUFFERPOOL size=16K,start_memory=12gb,memory=16gb

    Thanks,
    Hugo



    ------------------------------
    Hugo Tellez
    ------------------------------


  • 2.  RE: SHMTOTAL maxed on Workgroup Edition

    Posted Tue April 18, 2023 06:57 PM

    By those four BUFFERPOOL settings, you're allowing all of them to grow, to a combined maximum size of 30GB, with (initial) V segment already consuming 2GB.  If all bufferpools grew to their configured maximum (they currently add up to 29.6044 GB, plus some R and M memory), there'd be nothing left for additional virtual (V) memory.  New V segments often get allocated (even with seemingly sufficient free blocks in existing V segments) when a larger contiguous block of memory is needed (in your backup case: some 600kB).

    So, assuming you really need all four bufferpools (i.e. you have dbspaces utilizing the respective page sizes), you still shouldn't allow them to grow so big, or you'd switch to static configuration altogether.  A bufferpool cannot shrink, afaik, nor get deallocated.

     Andreas



    ------------------------------
    Andreas Legner
    ------------------------------



  • 3.  RE: SHMTOTAL maxed on Workgroup Edition

    Posted Tue April 18, 2023 08:40 PM

    I modified the bufferpools and restarted the instance.

    All good now.

    Thanks,
    Hugo



    ------------------------------
    Hugo Tellez
    ------------------------------