With glusterfs 1.3 getting close to a releasable state, I started thinking
about the structure of a global namespace for our applications.
There will be "mirrored machines" (AFR) for crucial data, and there should
be part of the servers offering read-only stuff (making sure nothing gets
added to th
Me again :-(
While experimenting with additional translators like read-ahead and
write-behind,
I found that each time I want to activate/deactivate a translator layer I have
to change the reference in the upper layer of the stack. This will sooner or
later result in typing errors, making the whol
One on my clients crashed early this morning. I believe it was idle at the
time. The bt:
#0 unify_sh_readdir_cbk (frame=0x14774670, cookie=0x80563b0, this=0x80568e0,
op_ret=0, op_errno=0, entry=0x14286fa0, count=459)
at unify-self-heal.c:283
283 list_for_each_entry (ino_list, list
I am trying to delete the contents of a tmp dir with 3 trees containing about
1.7G
as root, from withint the top level tmp dir I issue
rm -rf *
and the command hangs are never returns.
find . -exec rm -rf {} \;
works as expected.
Harris
___
You would get a 'connection refused' if the server is not running. Can you
please check if glusterfsd was running at that moment? also please get the
logs of the glusterfsd which was not running (and if possible, the core
dump's backtrace)
Also have you upgraded all the servers and clients?
than
Strange,
what is your configuration? At the time of 'hang', is it possible for you
to attach gdb to glusterfs and get a backtrace (from every thread, by
switching as 'thr 1' 'thr 2' etc) ?
rm -rf seems to work fine for me, wondering how find . -exec rm would make a
difference.
thanks,
avati
I am
the rm -rf hangs. It looks like one or two unlinks are sent to the log. I can
cntl-C the client and the data is still there. The data was is the tmp dir from
failed backups. It is gone now. I will investigate more when I have more data
later today.
Harris
- Original Message -
From: "An
On Thu, Jun 28, 2007 at 02:00:43PM +0200, Steffen Grunewald wrote:
> Me again :-(
>
> While experimenting with additional translators like read-ahead and
> write-behind,
> I found that each time I want to activate/deactivate a translator layer I have
> to change the reference in the upper layer o
Found another issue in the pre5.1 release...
In the doc/examples Makefile.am, there's a reference to $(docdir) which
under Debian would be mapped to /usr/share/doc/glusterfs (or similar).
Unfortunately, the configure script does *not* accept a --docdir= setting
(if I checkout branch2.4, and run aut
On Thu, Jun 28, 2007 at 04:41:02PM +0200, Steffen Grunewald wrote:
> Found another issue in the pre5.1 release...
And another one: the performance/stat-prefetch xlator has been replaced (?)
by performance/io-cache but the translators documentation doesn't reflect
that. I suppose the option(s) may
On Thu, Jun 28, 2007 at 07:31:26AM -0700, Vikas Gorur wrote:
> > Would it be possible to flag a translator as INACTIVE (kind of hand-through)
> > or ACTIVE (this should be the default) while keeping the whole stack
> > unchanged?
>
> This would be a nice feature to have. However, I'd vote for kee
Yes. After I sent this message, I realized that I neglected to upgrade the
client. However, after I upgraded the client and updated the config to include
the namespace info, the servers kept crashing. Since this was a production
machine, I had to downgrade as my "maintenance window" was over.
This happened when the first client in the client config
rebooted. Normally, the second client in the afr group would have picked up
the slack, but instead I was getting connection refused from the client. I
am assuming this is a locking issue?
You would get a 'connection refused' if the serv
Here is the debug-backtrace from glusterfsd right during/after it dies:
2007-06-28 14:59:57 E [protocol.c:262:gf_block_unserialize_transport]
libglusterfs/protocol: full_read of header failed: peer (127.0.0.1)
2007-06-28 14:59:57 C [tcp.c:81:tcp_disconnect] server: connection disconnected
2007-06
Majied,
This looks like the same bug in io-threads that I have already reported. Try
turning off io-threads and see if your problem goes away.
Harris
- Original Message -
From: "Majied Najjar" <[EMAIL PROTECTED]>
To: gluster-devel@nongnu.org
Sent: Thursday, June 28, 2007 3:07:20 PM (GM
Vikas Gorur wrote:
On Thu, Jun 28, 2007 at 02:00:43PM +0200, Steffen Grunewald wrote:
Me again :-(
While experimenting with additional translators like read-ahead and
write-behind,
I found that each time I want to activate/deactivate a translator layer I have
to change the reference in th
Here is the backtrace without io-threads, the same behavior:
2007-06-28 17:52:39 C [common-utils.c:205:gf_print_trace] debug-backtrace: Got
signal (6), printing backtrace
2007-06-28 17:52:39 C [common-utils.c:207:gf_print_trace] debug-backtrace:
/usr/lib/libglusterfs.so.0(gf_print_trace+0x2e) [0
Majied,
If you have a core file run:
gdb /usr/sbin/glusterfsd -c core.??? where core.??? is the name of the core file
then issue a bt the post the resulting backtrace. That will show the devs where
the crash is happening
Harris
- Original Message -
From: "Majied Najjar" <[EMAIL PROTEC
Easy enough:
This GDB was configured as "i486-linux-gnu".
(no debugging symbols found)
Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1".
Core was generated by `[glusterfsd]
'.
Program terminated with signal 11,
Howdy all,
I would like to request a feature which maybe quite easy with the
existing fixed-id translator.
The theory we have here is that we want to be able to have all our
servers use glusterfs (or a system like it) to store files, however at
the moment each server has a unique set of users and
20 matches
Mail list logo