[Gluster-users] Replace brick was crashed

[Thai. Ngo Bao]

Hello,

I was running a replace brick process in our cluster of 80TBs since last Friday. The process was running well and expected to finish in 6-7 days due to migration of lots of data. I, however, found that the replace brick process was crashed today. In addition, I issued the command to check the status of the process and it reported me that the migration was complete. Please have a look at the attachment for further info. 

I have tried to restart the process but did not have any success so far since a record of the status of this process was added in /etc/glusterd/vols/myvolume/info. 

Also, I found that the status of new brick, running the process, was "Peer Reject" in our cluster. I did try to peer detach and probe this brick again but always failed. 

Any thought? How can I restart the process?

Thanks,
~Thai
-------------- next part --------------
A non-text attachment was scrubbed...
Name: replace_brick_error.zip
Type: application/x-zip-compressed
Size: 893 bytes
Desc: replace_brick_error.zip
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20110830/0b725264/attachment.bin>