About a month ago, I upgraded my workstation to Nevada build 49.  When
I ran "zpool status" on my existing zfs pool (a couple mirrored
partitions, one from each of the system's two disks), I discovered that
the existing on-disk format was at version 1 and that I could upgrade
to version 3.  I chose to defer doing so, until I was sure that I would
no longer boot back into the partition holding my previous root (from
build 29).

Last week, I decided it was time to take the plunge and invoked
    % zpool upgrade zfs
The upgrade completed successfully, with no error messages.

But since that upgrade I've been getting unremitting bursts of disk
i/o.  When there's no activity on the system, they occur at a frequency
of 1/minute.  (With activity, the frequency is higher.  From listening
to my disks, there seems to be considerably more i/o traffic than
before the upgrade.)  Here's iostat info from when the system is
essentially idle:

% iostat 20
   tty        sd0           sd1           sd3           nfs1           cpu
 tin tout kps tps serv  kps tps serv  kps tps serv  kps tps serv   us sy wt id
   0    3 273  29  120  280  29   93    0   0    0    0   0    9   26  6  0 69
   0   12  11   5   20   11   5   17    0   0    0    0   0    0    2  1  0 97
   0    4   0   0    0    1   1   20    0   0    0    0   0    0    2  1  0 97
   0    4   2   0    3    1   0    6    0   0    0    0   0    0    2  1  0 97
   0    4  11   5   20   11   5   18    0   0    0    0   0    0    1  1  0 98
   0    4   0   0    0    0   0    0    0   0    0    0   0    0    1  1  0 98
   0    4   0   0    0    0   0    0    0   0    0    0   0    0    1  1  0 98
   0    4  11   5   18   11   5   17    0   0    0    0   0    0    1  1  0 98
   0    4   0   0    0    0   0    0    0   0    0    0   0    0    1  1  0 98
   0    4   0   0    0    0   0    0    0   0    0    0   0    0    1  1  0 98
   0    4  11   5   19   11   5   17    0   0    0    0   0    0    1  1  0 98
^C% 

This periodic activity did not occur before I upgraded my zfs pool, and
I can't think of any other change that might account for it.  The pool
is completely healthy:

% zpool status zfs
  pool: zfs
 state: ONLINE
 scrub: scrub completed with 0 errors on Fri Jan 12 15:34:05 2007
config:

        NAME          STATE     READ WRITE CKSUM
        zfs           ONLINE       0     0     0
          mirror      ONLINE       0     0     0
            c0t1d0s6  ONLINE       0     0     0
            c0t0d0s6  ONLINE       0     0     0

errors: No known data errors
% 

Have there been changes to the zfs implementation between builds 29 and
49 that a pool upgrade would trigger to cause periodic disk activity?
The noise from my disks is slowly driving me crazy (it's like Chinese
water torture...).

        -- Glenn Skinner

_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

Reply via email to