Sun 19 Aug 2007 11:14:58 PM UTC, original submission:
While leaving my servers over the weekend, with a rsync from hdd to glusterfs (to keep them uptodate) the glusterfs process died regually.
A core dump shows the following backtrace:
Core was generated by `[glusterfs] '.
Program terminated with signal 11, Segmentation fault.
#0 0xb7dd8abf in memcpy () from /lib/libc.so.6
(gdb) bt
#0 0xb7dd8abf in memcpy () from /lib/libc.so.6
#1 0xb7547ef7 in afr_lookup_cbk (frame=0xb69b3360, cookie=0xb5625300, this=0x805a7c0, op_ret=0, op_errno=2, inode=0x809c0a8, buf=0x0)
at ../../../../../xlators/cluster/afr/src/afr.c:436
#2 0xb75549ec in client_unlink_cbk (frame=0xb5625300, args=0xb4ec3410) at ../../../../../xlators/protocol/client/src/client-protocol.c:3207
#3 0xb7558513 in notify (this=0x805a340, event=2, data=0x809bf98) at ../../../../../xlators/protocol/client/src/client-protocol.c:4400
#4 0xb7eed807 in transport_notify (this=0x18, event=-1234241672) at ../../../libglusterfs/src/transport.c:154
#5 0xb7eee4a9 in sys_epoll_iteration (ctx=0xbfaaa470) at ../../../libglusterfs/src/epoll.c:53
#6 0xb7eed938 in poll_iteration (ctx=0xbfaaa470) at ../../../libglusterfs/src/transport.c:300
#7 0x0804a7fb in main (argc=4, argv=0xbfaaa554) at ../../../glusterfs-fuse/src/glusterfs.c:388
The glusterfs.log file shows:
2007-08-20 05:31:38 C [tcp.c:81:tcp_disconnect] hoggle-raw: connection disconnected
2007-08-20 05:31:38 E [afr.c:435:afr_lookup_cbk] ERROR: ../../../../../xlators/cluster/afr/src/afr.c: afr_lookup_cbk: (!buf) is true
---------
got signal (11), printing backtrace
---------
[0xb7ef8420]
/usr/local/lib/glusterfs/1.3.0/xlator/cluster/afr.so[0xb7547ef7]
/usr/local/lib/glusterfs/1.3.0/xlator/protocol/client.so[0xb75549ec]
/usr/local/lib/glusterfs/1.3.0/xlator/protocol/client.so(notify+0xaa3)[0xb7558513]
/usr/local/lib/libglusterfs.so.0(transport_notify+0x37)[0xb7eed807]
/usr/local/lib/libglusterfs.so.0(sys_epoll_iteration+0xd9)[0xb7eee4a9]
/usr/local/lib/libglusterfs.so.0(poll_iteration+0x78)[0xb7eed938]
[glusterfs](main+0x3fb)[0x804a7fb]
/lib/libc.so.6(__libc_start_main+0xdc)[0xb7d7eebc]
[glusterfs][0x804a371]
---------
The brick given (hoggle-raw) has been functioning correctly over the weekend and should never have been out of contact. (in fact its doing it now, even though the glusterfsd process is showing as running and traffic is being seen)
Unfortantly I am unable to determine what is causing the crash, as its done via an hourly cron job, but it should only be doing rsyncs between two data sources. (HDD => glusterfs)
|