[Gluster-users] Many problems in 3.1

Samuel Hassine sam at as.ht
Sat Oct 23 10:44:59 UTC 2010


Hello there,

I have just upgraded a GlusterFS filesystem in production after one week
of tests on a independant environment. First of all, I want to
congratulate Gluster team for the work and the new filesystem
capabilities, it is just awesome :)

But I am also very disappointed in the new internal replicate management
and the integrity checking. In production, we have a simple GlusterFS
system with 2 nodes in a replicate file system.

Since we migrate on Gluster 3.1 (from Gluster 3.0) with exactly the same
configuration, we pass from a daily partition log of 50k to 350M! We
have always so many errors like:

[2010-10-23 11:59:11.497818] W [fuse-bridge.c:1674:fuse_readv_cbk]
glusterfs-fuse: 56975335: READ => -1 (No such file or directory)
[2010-10-23 11:59:11.498753] W [fuse-bridge.c:1674:fuse_readv_cbk]
glusterfs-fuse: 56975336: READ => -1 (No such file or directory)
[2010-10-23 12:08:29.152259] W [fuse-bridge.c:1674:fuse_readv_cbk]
glusterfs-fuse: 57056236: READ => -1 (No such file or directory)
[2010-10-23 12:08:29.153387] W [fuse-bridge.c:1674:fuse_readv_cbk]
glusterfs-fuse: 57056237: READ => -1 (No such file or directory)
[2010-10-23 12:15:50.933772] W [fuse-bridge.c:1674:fuse_readv_cbk]
glusterfs-fuse: 57112661: READ => -1 (No such file or directory)

or like this:


 [2010-10-21 00:05:41.180371] W [fuse-bridge.c:1748:fuse_writev_cbk]
glusterfs-fuse: 36601886: WRITE => -1 (Permission denied)
[2010-10-21 00:05:41.974738] W [fuse-bridge.c:1748:fuse_writev_cbk]
glusterfs-fuse: 36601993: WRITE => -1 (Permission denied)
[2010-10-21 00:05:42.289002] W [fuse-bridge.c:1748:fuse_writev_cbk]
glusterfs-fuse: 36602074: WRITE => -1 (Permission denied)
[2010-10-21 00:05:43.79197] W [fuse-bridge.c:1748:fuse_writev_cbk]
glusterfs-fuse: 36602190: WRITE => -1 (Permission denied)
[2010-10-21 00:05:43.302434] W [fuse-bridge.c:1748:fuse_writev_cbk]

or like:

[2010-10-21 19:21:54.884030] I [afr-common.c:716:afr_lookup_done]
dns-replicate-0: background  meta-data data entry self-heal triggered.
path: ******/images/toolbar/icon-32-delete.png
[2010-10-21 19:21:54.889428] I
[afr-self-heal-common.c:1526:afr_self_heal_completion_cbk]
dns-replicate-0: background  meta-data data entry self-heal completed on
******/images/toolbar/icon-32-delete.png
[2010-10-21 19:21:54.917167] I [afr-common.c:662:afr_lookup_done]
dns-replicate-0: entries are missing in lookup of
*******/images/toolbar/icon-32-new.png.
[2010-10-21 19:21:54.917215] I [afr-common.c:716:afr_lookup_done]
dns-replicate-0: background  meta-data data entry self-heal triggered.
path: ********/images/toolbar/icon-32-new.png
[2010-10-21 19:21:54.922384] I
[afr-self-heal-common.c:1526:afr_self_heal_completion_cbk]
dns-replicate-0: background  meta-data data entry self-heal completed on
*******/images/toolbar/icon-32-new.png
[2010-10-21 19:21:54.936605] I [afr-common.c:662:afr_lookup_done]
dns-replicate-0: entries are missing in lookup of
******/images/toolbar/icon-32-new.png.

So our filesystem is unusable...

Somebody have the same problem? Is there a solution?

Thanks for your answer.

Regards.
Sam





More information about the Gluster-users mailing list