[Gluster-devel] GlusterFS 1.3.0-pre2.3: robustness

Anand Babu Periasamy ab at gnu.org.in
Thu Apr 5 01:27:58 UTC 2007

,----[ Gerry Reno writes: ]
| Ok, I setup a four brick AFR test and it is working.  Well at least
| as long as nothing goes wrong.  But if I take down one of the
| servers (glusterfsd) and try to write to any of the mounts then the
| whole thing hangs.  This is the type of situation that I am hoping
| will be handled gracefully.  I've had tcp/ip stacks hang, servers
| hang and need rebooted and once back up the whole cluster needs to
| be able to recover itself.

Gracefully handling connection timeouts and retries will be in pre3
release. We are hoping to make pre3 as the last preview release before
1.3 stable. 

1.4 will have Self-heal functionality with journaled, lazy recovery
mechanisms that will make AFR very smooth. As of now if a brick goes
down and you will have to use tools like rsync or unison to sync
before you bring it back online.

Anand Babu 
GPG Key ID: 0x62E15A31
Blog [http://ab.freeshell.org]              
The GNU Operating System [http://www.gnu.org]  

More information about the Gluster-devel mailing list