[OpenIndiana-discuss] Problems with zpool stalling

[Hans Joergensen]

Hello,

I've been having this problem with several storage servers running
mostly NFS clients.. (ESXi). And I've seen it both on nexenta and
OI.

I/O seems to freeze/pause when write IOPS are high..

Today I noticed it while having the zilstat.ksh and iostat running;

zilstat.ksh
   N-Bytes  N-Bytes/s N-Max-Rate    B-Bytes  B-Bytes/s B-Max-Rate ops  <=4kB 4-32kB >=32kB
  56030032   56030032   56030032  100241408  100241408  100241408 820      0     30    790
   2645352    2645352    2645352    5058560    5058560    5058560 50 0     11     39
     68792      68792      68792     131072     131072     131072 1 0      0      1
         0          0          0          0          0          0 0 0      0      0
         0          0          0          0          0          0 0 0      0      0
   6648152    6648152    6648152   12996608   12996608   12996608 125      0     19    106
     10576      10576      10576     708608     708608     708608 11 0      3      8
    663648     663648     663648    2539520    2539520    2539520 30 0      3     27
  58724208   58724208   58732968  109330432  109330432  109330432 916      5     16    896

iostat -xnz 5 showed everything - including busy - drop to around zero
at the same time.

I can't really figure out how to go from here... ? I think it's
weird if it's a ZIL-problem since the ZIL-devices are not busy when
the problem occurs.

Also I've seen this on multiple machines connected to different
switches.

Any ideas? Anyone seen something like this before?

btw, dmesg shows nothing.. And the number of SAS errors are
minimal..

Thanks,

Hans Jørgensen
Denmark.