Description of problem:
I'm sharing a volume between two web server.
After a problem on one brick, the volume was unreachable ((107)Transport endpoint is not connected) on the other brick too. Impossible to umount/mount/restart glusterfsd, i had to reboot the two server. This happened tree time this week (after more than two years running perfectly).
For the history, I was used to mount partitions using "glusterfs" as fstype. Four months ago, I switched to "nfs" as fstype for mounting partition, but we discovered a lot of NFS Stale files during last weeks, so, we switched back to "glusterfs" this week. Since we switched, one server (always the same) has crashed tree times.
For me crash can happen, but entire volume hang is a problem.


Additionnal informations :
Volume Name: ApacheRoot
Type: Replicate
Status: Started
Number of Bricks: 2
Transport-type: tcp
Bricks:
Brick1: xxxxx-srv-web06:/glusterfs/ApacheRoot
Brick2: xxxxx-srv-web05:/glusterfs/ApacheRoot
Options Reconfigured:
performance.io-thread-count: 32
performance.write-behind-window-size: 8MB
performance.cache-refresh-timeout: 3
performance.cache-max-file-size: 2MB
performance.cache-size: 1GB

df output :
xxxxx-srv-web06:/ApacheRoot      92G   17G   71G  19% /var/www


Here is what I've found in logs :

[2014-05-06 18:37:00.363251] W [afr-common.c:1121:afr_conflicting_iattrs] 0-ApacheRoot-replicate-0: /html/yyyown/cms/agents/cbnbdb/prod/_library/photo_gallery/kpa-map.jpg: gfid differs on subvolume 1 (26e0651f-8397-459d-b8cb-0657df085925, e3e846f9-0165-41bd-9f6e-e22825705653)
[2014-05-06 18:37:00.363278] E [afr-self-heal-common.c:1333:afr_sh_common_lookup_cbk] 0-ApacheRoot-replicate-0: Conflicting entries for /html/yyyown/cms/agents/cbnbdb/prod/_library/photo_gallery/kpa-map.jpg
[2014-05-06 18:37:00.364115] E [afr-self-heal-common.c:2074:afr_self_heal_completion_cbk] 0-ApacheRoot-replicate-0: background  entry self-heal failed on /html/yyyown/cms/agents/cbnbdb/prod/_library/photo_gallery
[2014-05-06 18:37:00.465741] E [client3_1-fops.c:2228:client3_1_lookup_cbk] 0-ApacheRoot-client-0: remote operation failed: Stale NFS file handle
[2014-05-06 18:37:00.466103] W [stat-prefetch.c:1549:sp_open_helper] 0-ApacheRoot-stat-prefetch: lookup-behind has failed for path (/html/yyyown/cms/agents/cbnbdb/prod/_library/photo_gallery/kpa-map.jpg)(Stale NFS file handle), unwinding open call waiting on it
[2014-05-06 18:37:00.466155] W [fuse-bridge.c:588:fuse_fd_cbk] 0-glusterfs-fuse: 34582930: OPEN() /html/yyyown/cms/agents/cbnbdb/prod/_library/photo_gallery/kpa-map.jpg => -1 (Stale NFS file handle)
[2014-05-06 18:37:00.467172] E [client3_1-fops.c:366:client3_1_open_cbk] 0-ApacheRoot-client-0: remote operation failed: No such file or directory
pending frames:
frame : type(1) op(OPEN)
frame : type(1) op(OPEN)

patchset: git://git.gluster.com/glusterfs.git
signal received: 11
time of crash: 2014-05-06 18:37:00
configuration details:
argp 1
backtrace 1
dlfcn 1
fdatasync 1
libpthread 1
llistxattr 1
setfsid 1
spinlock 1
epoll.h 1
xattr.h 1
st_atim.tv_nsec 1
package-string: glusterfs 3.2.5
/lib/x86_64-linux-gnu/libc.so.6(+0x364a0)[0x7f9693c004a0]
/usr/lib/glusterfs/3.2.5/xlator/performance/io-cache.so(ioc_open_cbk+0x99)[0x7f9690883af9]
/usr/lib/glusterfs/3.2.5/xlator/performance/read-ahead.so(ra_open_cbk+0x1ac)[0x7f9690a93bcc]
/usr/lib/glusterfs/3.2.5/xlator/performance/write-behind.so(wb_open_cbk+0x127)[0x7f9690c9fa17]
/usr/lib/glusterfs/3.2.5/xlator/cluster/replicate.so(afr_open_cbk+0x25e)[0x7f9690ecc6be]
/usr/lib/glusterfs/3.2.5/xlator/protocol/client.so(client3_1_open_cbk+0x228)[0x7f969111e188]
/usr/lib/libgfrpc.so.0(rpc_clnt_handle_reply+0xa5)[0x7f96943caec5]
/usr/lib/libgfrpc.so.0(rpc_clnt_notify+0x7d)[0x7f96943cb84d]
/usr/lib/libgfrpc.so.0(rpc_transport_notify+0x27)[0x7f96943c7ab7]
/usr/lib/glusterfs/3.2.5/rpc-transport/socket.so(socket_event_poll_in+0x34)[0x7f9691d58074]
/usr/lib/glusterfs/3.2.5/rpc-transport/socket.so(socket_event_handler+0xc7)[0x7f9691d583c7]
/usr/lib/libglusterfs.so.0(+0x3bce7)[0x7f969460ece7]
/usr/sbin/glusterfs(main+0x2a5)[0x7f9694a5e4b5]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xed)[0x7f9693beb76d]
/usr/sbin/glusterfs(+0x4745)[0x7f9694a5e745]


Version-Release number of selected component (if applicable):
ubuntu 12.10 LTS
glusterfs 3.2.5




How reproducible:
Dunno

Steps to Reproduce:
1.
2.
3.

Actual results:
Volume hang

Expected results:
No volume hang

Additional info: