I'm seeing some strange behavior on my new hardware and would like to
figure out some way to diagnose what's going on.  It's not a serious
problem yet, but I'd like to know if this will be indicative of things
I'll see later.

I have a 7-disk striped zpool composed of whatever I had lying around
(i.e., I don't really trust these disks).  I'm writing large
sequential blocks to it with dd.  Watching in 'zpool iostat -v scratch
10', I see things like this (summarized; full output later):

scratch      391G  1.90T      0  2.44K  12.8K   308M
scratch      393G  1.90T      0  1.23K  57.5K   156M
scratch      393G  1.90T      0      0      0      0 (does this a few times)
scratch      396G  1.89T      0  1.76K  6.39K   220M

How can I diagnose which disk (or other piece of hardware) is causing
these slowdowns?  This system isn't in production, so I can do
arbitrary things to it.

The (almost) full output of zpool iostat follows:

               capacity     operations    bandwidth
pool         used  avail   read  write   read  write
----------  -----  -----  -----  -----  -----  -----
scratch      391G  1.90T      0  2.44K  12.8K   308M
  c4t9d0    67.2G   629G      0    432      0  53.5M
  c4t10d0   54.0G  94.0G      0    339  6.39K  42.2M
  c4t11d0   54.3G   244G      0    332      0  41.5M
  c4t12d0   54.0G   244G      0    347      0  42.8M
  c4t13d0   54.0G   244G      0    350      0  43.0M
  c4t14d0   54.0G   244G      0    349      0  42.6M
  c4t15d0   54.0G   244G      0    345  6.39K  42.3M
----------  -----  -----  -----  -----  -----  -----

               capacity     operations    bandwidth
pool         used  avail   read  write   read  write
----------  -----  -----  -----  -----  -----  -----
scratch      393G  1.90T      0  1.23K  57.5K   156M
  c4t9d0    67.4G   629G      0    190  6.39K  23.5M
  c4t10d0   54.2G  93.8G      0    172      0  21.5M
  c4t11d0   54.5G   243G      0    194      0  24.3M
  c4t12d0   54.2G   244G      0    170  19.2K  21.3M
  c4t13d0   54.2G   244G      0    175  12.8K  21.6M
  c4t14d0   54.2G   244G      0    178  12.8K  21.7M
  c4t15d0   54.2G   244G      0    179  6.39K  21.8M
----------  -----  -----  -----  -----  -----  -----

               capacity     operations    bandwidth
pool         used  avail   read  write   read  write
----------  -----  -----  -----  -----  -----  -----
scratch      393G  1.90T      0      0      0      0
(snipped here; all disks are zeroes all the way across)
scratch      393G  1.90T      0      0      0      0
scratch      393G  1.90T      0      0      0      0
scratch      393G  1.90T      0      0      0      0
scratch      393G  1.90T      0      0      0      0
scratch      393G  1.90T      0      0      0      0
----------  -----  -----  -----  -----  -----  -----

               capacity     operations    bandwidth
pool         used  avail   read  write   read  write
----------  -----  -----  -----  -----  -----  -----
scratch      396G  1.89T      0  1.76K  6.39K   220M
  c4t9d0    68.0G   628G      0    328      0  40.8M
  c4t10d0   54.6G  93.4G      0    239      0  29.3M
  c4t11d0   54.9G   243G      0    244      0  30.2M
  c4t12d0   54.6G   243G      0    245      0  29.7M
  c4t13d0   54.6G   243G      0    249      0  30.1M
  c4t14d0   54.7G   243G      0    249      0  30.3M
  c4t15d0   54.6G   243G      0    241  6.39K  29.8M
----------  -----  -----  -----  -----  -----  -----

Will
_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to