[Gluster-devel] AFR setup with Virtual Servers crashes

[Urban Loesch]

Hi Avati,

I tried out the latest code from repository.
But I'm not sure if I made all correct because TLA is new for me.

Downloaded the latest source:
- # tla get -A gluster at sv.gnu.org glusterfs--mainline--2.5 glusterfs
- executed "autogen.sh" without errors
- # ./configure the Server and Client
    - Server: CFLAGS="-O3" ./configure --prefix=/usr --sysconfdir=/etc 
--disable-fuse-client --disable-ibverbs
    - Client: CFLAGS="-O3" ./configure --prefix=/usr --sysconfdir=/etc 
--disable-server --disable-ibverbs
- # make
- # make install

Now "glusterfsd -V" shows me the following at both servers.
glusterfs 1.4.0
Copyright (c) 2006, 2007 Z RESEARCH Inc. <http://www.zresearch.com>
GlusterFS comes with ABSOLUTELY NO WARRANTY.
You may redistribute copies of GlusterFS under the terms of the GNU 
General Public License.

And "glusterfs -V" as follows:
glusterfs 1.4.0
Copyright (c) 2006, 2007 Z RESEARCH Inc. <http://www.zresearch.com>
GlusterFS comes with ABSOLUTELY NO WARRANTY.
You may redistribute copies of GlusterFS under the terms of the GNU 
General Public License.

The both servers are working normally and I can mount the volumes on the 
client without problems, but when I try some directory listing (eg. ls 
-lh mounted directory) it shows me the content of the directory and the 
the mounted volume crashes now at the client. The servers are working 
nomally

It happens any time.
Here's the core dump of the crash:
gdb glusterfs -c core.3701
..
Failed to read a valid object file image from memory.
Core was generated by `glusterfs --no-daemon --log-file=/dev/stdout 
--log-level=DEBUG --spec-file=/etc'.
Program terminated with signal 11, Segmentation fault.
#0  0xb7f1a016 in dict_get (this=0x8077050, key=0x8055960 "brick") at 
dict.c:132
132     dict.c: No such file or directory.
        in dict.c
(gdb) bt
#0  0xb7f1a016 in dict_get (this=0x8077050, key=0x8055960 "brick") at 
dict.c:132
#1  0xb75a8920 in client_releasedir () from 
/usr/lib/glusterfs/1.4.0/xlator/protocol/client.so
#2  0xb75a24fe in afr_releasedir () from 
/usr/lib/glusterfs/1.4.0/xlator/cluster/afr.so
#3  0x08050ac6 in fuse_releasedir ()
#4  0xb7f0e73f in fuse_reply_err () from /usr/lib/libfuse.so.2
#5  0xb7f0fa8d in fuse_reply_entry () from /usr/lib/libfuse.so.2
#6  0xb7f112d6 in fuse_session_process () from /usr/lib/libfuse.so.2
#7  0x0804a8e5 in fuse_transport_notify ()
#8  0xb7f210fd in transport_notify (this=0x80553c0, event=1) at 
transport.c:148
#9  0xb7f21da9 in sys_epoll_iteration (ctx=0xbfdd3a08) at epoll.c:53
#10 0xb7f211ad in poll_iteration (ctx=0xbfdd3a08) at transport.c:251
#11 0x0804a1db in main ()
(gdb) quit

Have you any idea or have I made some error during compilation?

thanks
Urban


Anand Avati wrote:
> Urban,
>  this bug has alredy been fixed in the source repository.
> thanks,
> avati
>
> 2007/5/10, Urban Loesch <ul at enas.net>:
>> Hi Avati,
>>
>> thanks for your fast answer.
>>
>> I use the version glusterfs-1.3.0-pre3 downloaded form your server
>> (http://ftp.zresearch.com/pub/gluster/glusterfs/1.3-pre/).
>> I will try the latest version from TLA today afternoon and let you know
>> what happens.
>>
>> Here's the backtrace from the core dump
>> # gdb glusterfsd -c core.15160
>> ..
>> Core was generated by `glusterfsd --no-daemon --log-file=/dev/stdout
>> --log-level=DEBUG'.
>> Program terminated with signal 11, Segmentation fault.
>> #0  0xb75d8fd3 in posix_locks_flush () from
>> /usr/lib/glusterfs/1.3.0-pre3/xlator/features/posix-locks.so
>> (gdb) bt
>> #0  0xb75d8fd3 in posix_locks_flush () from
>> /usr/lib/glusterfs/1.3.0-pre3/xlator/features/posix-locks.so
>> #1  0xb75d1192 in fop_flush () from
>> /usr/lib/glusterfs/1.3.0-pre3/xlator/protocol/server.so
>> #2  0xb75cded7 in proto_srv_notify () from
>> /usr/lib/glusterfs/1.3.0-pre3/xlator/protocol/server.so
>> #3  0xb7f54ecd in transport_notify (this=0x804b1a0, event=1) at
>> transport.c:148
>> #4  0xb7f55b79 in sys_epoll_iteration (ctx=0xbfbc2ff0) at epoll.c:53
>> #5  0xb7f54f7d in poll_iteration (ctx=0xbfbc2ff0) at transport.c:251
>> #6  0x0804924e in main ()
>>
>> Yes it is reproducible. It happens every time when I try to start my
>> virtual server.
>>
>> Thanks
>> Urban
>>
>
>