[Gluster-devel] The return of the all-null pending matrix

Vijay Bellur vbellur at redhat.com
Sun Jul 14 19:19:16 UTC 2013


On 07/15/2013 12:41 AM, Emmanuel Dreyfus wrote:
> Vijay Bellur <vbellur at redhat.com> wrote:
>
>> I checked the logs and did not notice anything unusual in the logs.
>> Do you have multiple clients involved? Does turning off eager-locking
>> through:
>
> No multiple clients. Is it then relevant to test eager-lock?
>

FWIW, it is. self-heal daemon and a client can also contend for locks. 
Messages of this sort make me wonder where the contention is coming from:

[2013-06-28 21:45:30.217139] I [afr-lk-common.c:1075:afr_lock_blocking] 
0-gfs34-replicate-0: unable to lock on even one child
[2013-06-28 21:45:30.217189] I 
[afr-transaction.c:1063:afr_post_blocking_inodelk_cbk] 
0-gfs34-replicate-0: Blocking inodelks failed.

-Vijay




More information about the Gluster-devel mailing list