[Gluster-devel] Bug report: Seg. Fault on unify when re-running dbench (patch 284)

Amar S. Tumballi amar at zresearch.com
Sun Jul 8 09:05:15 UTC 2007


Hi Daniel,
  If you still have the core, can you mail me back with the following
things?

# gdb glusterfs -c /core.<pid>
.
.
.
(gdb) p *loc->inode
.
(gdb) p *list
.
(gdb) p *ino_list


Thanks and regards,
Amar

On 7/8/07, Daniel van Ham Colchete <daniel.colchete at gmail.com> wrote:
>
> Hello all!
>
> I was doing some tests with GlusterFS here and I think I found a bug.
>
> Current setup: 2 servers (192.168.0.151 and 192.168.0.155) and 1 client (
> 192.168.0.155).
> How to reproduce: run "dbench -t 10 -c /usr/share/dbench/client.txt 50",
> CTRL+C after a few seconds, run "dbench -t 10 -c
> /usr/share/dbench/client.txt 50" again.
>
>
> Client spec file:
> volume brick1
>         type protocol/client
>         option transport-type tcp/client
>         option remote-host 192.168.0.151
>         option remote-port 6991
>         option remote-subvolume brick
> end-volume
>
> volume brick2
>         type protocol/client
>         option transport-type tcp/client
>         option remote-host 192.168.0.155
>         option remote-port 6991
>         option remote-subvolume brick
> end-volume
>
> volume brick-ns
>         type protocol/client
>         option transport-type tcp/client
>         option remote-host 192.168.0.151
>         option remote-port 6992
>         option remote-subvolume brick
> end-volume
>
> volume unify
>         type cluster/unify
>         subvolumes brick1 brick2
>         option namespace brick-ns
>         option scheduler rr
>         option rr.limits.min-free-disk 5
> end-volume
>
> volume writebehind
>         type performance/write-behind
>         option aggregate-size 131072
>         subvolumes unify
> end-volume
>
>
> Server-151.spec:
> volume brick
>         type storage/posix
>         option directory /mnt/cfs-01
> end-volume
>
> volume server
>         type protocol/server
>         option transport-type tcp/server
>         option listen-port 6991
>         subvolumes brick
>         option auth.ip.brick.allow 192.168.*
> end-volume
>
>
>
> Server-ns.spec:
> volume brick
>         type storage/posix
>         option directory /mnt/cfs-ns
> end-volume
>
> volume server
>         type protocol/server
>         option transport-type tcp/server
>         option listen-port 6992
>         subvolumes brick
>         option auth.ip.brick.allow 192.168.*
> end-volume
>
>
>
> Server-155.spec
> volume brick
>         type storage/posix
>         option directory /mnt/cfs-01
> end-volume
>
> volume server
>         type protocol/server
>         option transport-type tcp/server
>         option listen-port 6991
>         subvolumes brick
>         option auth.ip.brick.allow 192.168.*
> end-volume
>
>
>
> Backtrace:
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread -1210407248 (LWP 6202)]
> 0xb759102a in unify_opendir (frame=0x80c29e8, this=0x80587f0,
> loc=0x809b50c)
> at unify.c:1730
> 1730      list_for_each_entry (ino_list, list, list_head)
> (gdb) where
> #0  0xb759102a in unify_opendir (frame=0x80c29e8, this=0x80587f0,
> loc=0x809b50c) at unify.c:1730
> #1  0xb7f13de9 in default_opendir (frame=0x80d03d0, this=0x8058e48,
> loc=0x809b50c) at defaults.c:872
> #2  0x0804e7fb in fuse_opendir (req=0x83137d8, ino=8388904, fi=0xbfd7df2c)
> at fuse-bridge.c:1225
> #3  0xb7ef7b1c in ?? () from /usr/lib/libfuse.so.2
> #4  0x083137d8 in ?? ()
> #5  0x00800128 in ?? ()
> #6  0xbfd7df2c in ?? ()
> #7  0x00010000 in ?? ()
> #8  0x00000000 in ?? ()
>
>
> Last lines from the client log:
> 2007-07-07 20:24:37 D [fuse-bridge.c:441:fuse_getattr] glusterfs-fuse:
> GETATTR 16818308 (/clients/client37/~dmtmp/WORD)
> 2007-07-07 20:24:37 D [fuse-bridge.c:441:fuse_getattr] glusterfs-fuse:
> GETATTR 4194498 (/clients/client26)
> 2007-07-07 20:24:37 D [fuse-bridge.c:320:fuse_entry_cbk] glusterfs-fuse:
> ERR
> => -1 (2)
> 2007-07-07 20:24:37 D [inode.c:481:__create_inode] brick-ns/inode: create
> inode(8388904)
> 2007-07-07 20:24:37 D [inode.c:279:__active_inode] brick-ns/inode:
> activating inode(8388904), lru=0/1000
> 2007-07-07 20:24:37 D [inode.c:481:__create_inode] unify/inode: create
> inode(8388904)
> 2007-07-07 20:24:37 D [inode.c:279:__active_inode] unify/inode: activating
> inode(8388904), lru=0/1000
> 2007-07-07 20:24:37 D [inode.c:481:__create_inode] brick2/inode: create
> inode(2637863)
> 2007-07-07 20:24:37 D [inode.c:279:__active_inode] brick2/inode:
> activating
> inode(2637863), lru=0/1000
> 2007-07-07 20:24:37 D [fuse-bridge.c:441:fuse_getattr] glusterfs-fuse:
> GETATTR 29360322 (/clients/client29)
> 2007-07-07 20:24:37 D [inode.c:481:__create_inode] brick1/inode: create
> inode(7913510)
> 2007-07-07 20:24:37 D [inode.c:279:__active_inode] brick1/inode:
> activating
> inode(7913510), lru=0/1000
> 2007-07-07 20:24:37 D [fuse-bridge.c:291:fuse_entry_cbk] glusterfs-fuse:
> ENTRY => 8388904
> 2007-07-07 20:24:37 D [inode.c:310:__passive_inode] unify/inode:
> passivating
> inode(8388904), lru=1/1000
> 2007-07-07 20:24:37 D [fuse-bridge.c:441:fuse_getattr] glusterfs-fuse:
> GETATTR 25166149 (/clients/client2)
>
>
> Best regards,
> Daniel Colchete
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at nongnu.org
> http://lists.nongnu.org/mailman/listinfo/gluster-devel
>



-- 
Amar Tumballi
http://amar.80x25.org
[bulde on #gluster/irc.gnu.org]



More information about the Gluster-devel mailing list