xtrabackup failing due to log block number mismatch or child process dead

DBD::mysql::db selectrow_array failed: MySQL server has gone away at /usr/bin/innobackupex line 1325.

Having gone through the details, it was identified as a bug. Could you please suggest if this is taken into consideration and fixed in the latest releases if so share the release details so that I can upgrade xtracbakup for smooth functionality.

There are some bugs reported on it. You can try latest version of Percona Xtrabackup but before this it probably be related to redo log files size. Can you please try to increase the InnoDB redo log files size and retry when server is not heavily loaded because of write activity it not able to keep up with log copying thread. Also, can you please share your complete backup command with InnoDB specific parameters of my.cnf

Thanks for the update Irfan. Will try to share the my.cnf parameters and will try with the latest version of xtrabackup and get in touch with you.

Here are the DB related parameters in my.cnf
query_cache_type = 0
query_cache_size = 10M
Backup command is a pearl script, might not be able to share that due security reasons. Could you please suggest which parameter value to be changed and how much. Will try to install the latest xtracbackup 2.1 with the parameter value and post the status. Thanks again for your prompt response!

Hi Irfan,

Percona xtrabackup is upgraded to the latest version "percona-xtrabackup.x86_64

Required parameters are shared above. Kindly suggest a permanent solution please!

Best Regards,

Did you tried increasing innodb_log_file_size and try to backup during low database load. Look likes your database server is actively busy during backup.

Currently innodb_log_file_size in our case it is 2GB. How much do you want me to increase to! Load on the database should be quite low since backup runs during non-business hours and we do not run anyother jobs or processes on the slave host.