Galera Replication error: log block numbers mismatch

Question

I am attempting to get Galera replication working between two nodes.

I am finding these errors in my innobackup.backup.log file:

xtrabackup: error: log block numbers mismatch:
xtrabackup: error: expected log block no. 671400745, but got no. 679592737 from the log file.
xtrabackup: error: it looks like InnoDB log has wrapped around before xtrabackup could process all records due to either log copying being too slow, or log files being too small.
xtrabackup: Error: xtrabackup_copy_logfile() failed.

I am not sure where to do from here, I have googled a bit and didn't find anything that seemed like it could help me in my particular instance.

Any advice will be most appreciated.

Answer 1

I finally have solved this problem.

I found an article on launchpad that said to use this command to see if the problem was IO:
innobackupex --user=<username> --password=<password> --stream=tar -- ibbackup=/usr/bin/xtrabackup /tmp >/dev/null

This failed to, so I started thinking IO was not at fault.

I played around with this command and found this one to complete:
innobackupex --user=<username> --password=<password> --stream=xbstream --parallel=15 /tmp >/dev/null

This also showed that my ulimit was not high enough, so I had to set this higher for the backup to complete.

To force innobackupex to use xbstream instead of tar for the --stream option I but this in my.cnf on both of my galera servers:
[sst] streamfmt=xbstream