[Gluster-devel] [bug #19614] System crashes when node fails even with xfr

[Krishna Srinivas]

Hi,

If my understanding of your setup is correct,
your setup has four AFRs and each AFR replicates
to two bricks. When you kill one of the glusterfsds,
the client says "Transport end point not connected"
Is that correct?

It should not have happened. Can you paste your
config files exactly as you are using. Also which
glusterfsd did you kill?

Thanks
Krishna

On 4/17/07, anonymous <INVALID.NOREPLY at gnu.org> wrote:
>
> URL:
>   <http://savannah.nongnu.org/bugs/?19614>
>
>                  Summary: System crashes when node fails even with xfr
>                  Project: Gluster
>             Submitted by: None
>             Submitted on: Tuesday 04/17/2007 at 08:04 UTC
>                 Category: GlusterFS
>                 Severity: 3 - Normal
>                 Priority: 5 - Normal
>               Item Group: Crash
>                   Status: None
>                  Privacy: Public
>              Assigned to: None
>              Open/Closed: Open
>          Discussion Lock: Any
>         Operating System: GNU/Linux
>
>     _______________________________________________________
>
> Details:
>
> When testing I killed the glusterfsd process (and it also occured after a
> reboot of the server) a few moments after, read/write operations to the
> mountpoint started failing with 'Transport endpoint is not connected'
>
> after the node was restarted the system may recover, however this does not
> always happen.
>
> The docs say afr is RAID1, normally RAID1 will continue to operate when a
> drive is lost.
>
>
> Version: glusterfs-1.3.0-pre2.3
> OS: debain4.0 (etch) with fuse libs from unstable
> Appropriate Client Config Entries:
> volume server1-raw
>   type protocol/client
>   option transport-type tcp/client # for TCP/IP transport
>   option remote-host x.x.x.1 # IP address of the remote brick
>   option remote-subvolume bricklocal # name of the remote volume
> end-volume
>
> Specification Add client feature and attach to remote subvolume of server1
> volume server1-backup
>   type protocol/client
>   option transport-type tcp/client # for TCP/IP transport
>   option remote-host x.x.x.2 # IP address of the remote brick
>   option remote-subvolume brickbackup # name of the remote volume
> end-volume
> volume server1
>   type cluster/afr
>   subvolumes server1-raw server1-backup
>   option replicate *:2
> end-volume
> ... (duplicated for server2 thu 4) ...
>
> volume bricks
>   type cluster/unify
>   subvolumes server1 server2 server3 server4
>
>   option scheduler alu   # use the ALU scheduler
>   option alu.limits.min-free-disk  500MB                # Min free diskspace
> ...
>
>
>
>
>
>     _______________________________________________________
>
> Reply to this item at:
>
>   <http://savannah.nongnu.org/bugs/?19614>
>
> _______________________________________________
>   Message sent via/by Savannah
>   http://savannah.nongnu.org/
>
>
>
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at nongnu.org
> http://lists.nongnu.org/mailman/listinfo/gluster-devel
>