Comment 12 for bug 1240098

Hi,

I'm able to reproduce this on our percona cluster too ...
I run into a problem with "Deadlock found ..." while replicating from another outstanding server into my cluster (on node1).

After probing a while I did a search for this error and found this bug.

Our current setup is:

(master node) -> (cluster node1) <-> (cluster node2) <-> (cluster node3)
Means we have a 3 node percona cluster currently getting their updates only from a replication configured on node1.

While replicating data works fine for several days/weeks a few days ago I found this deadlock in the slave status output.

Restarting slave and even the cluster node did not change anything.
I also shut down all other nodes and tried to start the replication stand alone with the first node, but got the error in this case too.

After I found this bug report I tried the steps to reproduce the bug from comment #1 and was able to reproduce it.
I have installed version 5.5.34 of percona cluster.
All my nodes and the external master too runs with row based replication.

Since one told the error cannot be reproduced in version 5.5.31 I just downgraded to this version and really, there the example form #1 runs without error.

I'm just waiting for my replication to came up to the point of error to check, if the error still happens there.
Will inform you once I have checked this ...