zfs diff deadlock

[Peter Jeremy]

I recently decided to do a "zfs diff" between two snapshots to try
and identify why there was so much "USED" space in the snapshot.
The diff ran for a while (though with very little IO) but has now
wedged unkillably.  There's nothing on the console or in any logs,
the pool reports no problems and there are no other visible FS
issues.  Any ideas on tracking this down?

root at server:/root # ps -lp 53146
  UID   PID  PPID CPU PRI NI   VSZ   RSS MWCHAN STAT  TT       TIME COMMAND
    0 53146 53092   0  76  0 18096   948 ffffff0149cc6100 D+    33    4:18.76 zfs diff -F tank/beckett/home at 20121108 tank/beckett/home at 20121109
root at server:/root # procstat -kk 53146
  PID    TID COMM             TDNAME           KSTACK                       
53146 100490 zfs              -                mi_switch+0x18e sleepq_wait+0x44 _sx_slock_hard+0x22c _sx_slock+0x7a zap_get_leaf_byblk+0xbd zap_deref_leaf+0x68 fzap_cursor_retrieve+0xe7 zap_cursor_retrieve+0x155 zap_value_search+0x7f zfs_obj_to_path_impl+0x292 zfs_obj_to_stats+0x175 zfs_ioc_obj_to_stats+0x7e zfsdev_ioctl+0xe6 devfs_ioctl_f+0x7a kern_ioctl+0xb6 ioctl+0xf0 amd64_syscall+0x109 Xfast_syscall+0xfc 
53146 100643 zfs              initial thread   mi_switch+0x18e sleepq_catch_signals+0x247 sleepq_wait_sig+0x16 _sleep+0x2be pipe_write+0x6c8 write_record+0x79 diff_cb+0x166 traverse_visitbp+0x21c traverse_visitbp+0x316 traverse_visitbp+0x316 traverse_visitbp+0x316 traverse_visitbp+0x316 traverse_visitbp+0x316 traverse_visitbp+0x316 traverse_dnode+0x7c traverse_visitbp+0x48c traverse_impl+0x188 traverse_dataset+0x32 
root at server:/root # zpool status tank
  pool: tank
 state: ONLINE
  scan: resilvered 1.58G in 0h10m with 0 errors on Tue Oct 30 06:55:29 2012
config:

        NAME        STATE     READ WRITE CKSUM
        tank        ONLINE       0     0     0
          raidz2-0  ONLINE       0     0     0
            ada2p5  ONLINE       0     0     0
            ada3p5  ONLINE       0     0     0
            ada1p5  ONLINE       0     0     0
            ada5p5  ONLINE       0     0     0
            ada0p5  ONLINE       0     0     0
            ada4p5  ONLINE       0     0     0
        logs
          ada6p3    ONLINE       0     0     0
        cache
          ada6p4    ONLINE       0     0     0

errors: No known data errors

The systems is running a 4-month old 8-stable (r237444)

-- 
Peter Jeremy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
URL: <http://lists.freebsd.org/pipermail/freebsd-fs/attachments/20121111/dddde322/attachment.sig>