Hi Benjamin,
We have been running tests and production environments on redhat, centos,
ubuntu and slackware. There are many users (of whom I may not know), who've
been running Debian.
Can you send us the log files (both client and server) and backtrace of the
crash (use gdb command 'bt full' after running gdb on the core using
'gdb -c
<path to core file> <path to glusterfs binary>)?
regards,
On Fri, Jan 15, 2010 at 9:08 PM, Benjamin Hudgens <bhudgens at
photodex.com>wrote:
> Hello --
>
>
>
> Thus far we have made several attempts at bringing a stable gluster
> solution online. I 'need' Gluster to work for me. It is the only
> solution I have found that perfectly fits the mold of a file system we
> desperately need to implement (without code changes).
>
>
>
> Unfortunately, I have been unsuccessful at getting Gluster to stay
> stable. Most notably one-or-more of the client daemons will crash
> leaving Linux thinking it still has a mount point but when things access
> that mount point nobody is home.
>
>
>
> Before I invest additional time into troubleshooting I was hoping
> someone would indicate if they have a stable version running. If yes,
> which flavor of linux are you running on? We have been using Debian.
> Compile is clean yet still crashes. We have tried version 2.04, 2.08,
> 3.0, 3.0.1 (I think), or whatever this latest one was. I'd like to
> start with someone else's stable setup and go from there.
>
>
>
> For testing we are doing the following:
>
>
>
> Machine A runs a server distributing 3 drives
>
> Machine A runs a client accessing those 3 drives (Drive A-1)
>
> Machine B-E run servers with distributed/replicated drives
>
> Machine A runs a client accessing B-E (Drive A-2)
>
>
>
> I currently mount a samba share onto Machine A and run tons of rsyncs
> against the local client/server (Drive A-1). I then run that same rsync
> from Drive A-1 -> Drive A-2.
>
>
>
> Again, before any investment would be made sending debug
> information/troubleshooting... I was hoping someone would simply suggest
> the most stable scenario to run GlusterFS in. We will make adjustments
> accordingly and start from scratch.
>
>
>
> Thanks you for any assistance provided!
>
>
>
> Kind Regards,
>
> Benjamin
>
>
>
> ----------------------------------
>
> Benjamin Hudgens
>
> Director of Information Technology
>
> Photodex Corporation
>
>
>
> www.photodex.com <http://www.photodex.com/>
>
> bhudgens at photodex.com <mailto:bhudgens at photodex.com>
>
> (512) 674-9920
>
>
>
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
>
>
--
Raghavendra G