[DRBD-user] umount block when do resync operation

Mia Lueng xiaozunvlg at gmail.com
Tue Mar 19 08:07:23 CET 2013

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.

when I umount the drbd disk during the resync, it will take a long time to
do it and kernel oops occus like this:

[root at kvm03 /oradata]# cat /proc/drbd
version: 8.3.13 (api:88/proto:86-97)
 0: cs:SyncSource ro:Primary/Secondary ds:UpToDate/Inconsistent A r-----
    ns:13899984 nr:0 dw:3006320 dr:11798141 al:1426 bm:653 lo:2 pe:7133
ua:0 ap:2 ep:1 wo:d oos:2961452
[==============>.....] sync'ed: 77.3% (2892/12716)M
finish: 0:09:23 speed: 5,248 (4,300) K/sec
 1: cs:Unconfigured

here is kernel oops:

INFO: task umount:8135 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
umount        D 0000000000000003     0  8135   8134 0x00000080
 ffff881f51dcdc28 0000000000000086 ffff881f51dcdb88 ffffffff810e2f94
 ffff881f51dcdba8 ffffffff814f9392 0000000000000000 0000000000000000
 ffff881f7fc9e5f8 ffff881f51dcdfd8 000000000000f4e8 ffff881f7fc9e5f8
Call Trace:
 [<ffffffff810e2f94>] ? __delayacct_blkio_end+0x34/0x60
 [<ffffffff814f9392>] ? io_schedule+0x92/0xc0
 [<ffffffff814f9a6d>] schedule_timeout+0x22d/0x300
 [<ffffffff81065634>] ? enqueue_task_fair+0x64/0x100
 [<ffffffff81059c67>] ? enqueue_task+0x77/0x90
 [<ffffffff814f96d3>] wait_for_common+0x123/0x180
 [<ffffffff8105f4a0>] ? default_wake_function+0x0/0x20
 [<ffffffff814f97ed>] wait_for_completion+0x1d/0x20
 [<ffffffff811a6648>] sync_inodes_sb+0x88/0x1a0
 [<ffffffff811ac532>] __sync_filesystem+0x82/0x90
 [<ffffffff811ac73b>] sync_filesystem+0x4b/0x70
 [<ffffffff8117f387>] generic_shutdown_super+0x27/0xe0
 [<ffffffff8117f471>] kill_block_super+0x31/0x50
 [<ffffffff81180430>] deactivate_super+0x70/0x90
 [<ffffffff8119c12f>] mntput_no_expire+0xbf/0x120
 [<ffffffff8119cdcb>] sys_umount+0x25b/0x4b0
 [<ffffffff8100b0f2>] system_call_fastpath+0x16/0x1b

[root at kvm3 ~]# drbdadm dump drbd0
# resource drbd0 on kvm3.hgccp: not ignored, not stacked
resource drbd0 {
    protocol               A;
    on kvm3.hgccp {
        device           /dev/drbd0 minor 0;
        disk             /dev/vg_kvm3/drbd0;
        address          ipv4;
        meta-disk        internal;
    on kvm4.hgccp {
        device           /dev/drbd0 minor 0;
        disk             /dev/vg_kvm4/drbd0;
        address          ipv4;
        meta-disk        internal;
    net {
        ping-timeout      30;
        ping-int          30;
    disk {
        on-io-error      pass_on;
    syncer {
        rate              5M;
        csums-alg        md5;
        verify-alg       md5;
        c-max-rate       10M;
        c-min-rate         0;
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20130319/0784c732/attachment.htm>

More information about the drbd-user mailing list