[Gluster-users] split brain re-occuring

[Machiel Groeneveld]

Our set up is 2 replicated bricks with quorum set to 'auto'. No other
tuning has been done.

The problem is we're getting frequent I/O errors in zip files that are on
the dfs. One brick has an old version and the other the new file. The
client won't read the file if the servers don't agree on the file.

I ran the find command to self-heal the volume (running stat on all files),
fixed all the IO errors by removing the file from one brick, ran find again
and no errors were shown. But today the IO errors return.

Any hints on the cause of this issue? Or something I have to look at to
find out more?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20140411/3053761f/attachment.html>