We are having issues when trying to backup our Percona XtraDB Cluster 8.0.23 with Xtrabackup 8.
Our setup is as follows:
Two nodes (N1, N2) with Arbiter (A).
App is connecting only to the N1, N2 is only a failover and host for creating backups.
We are trying to create backup running this command on N2:
xtrabackup --ftwrl-wait-timeout=180 --lock-ddl-timeout=180 --backup-lock-timeout=180 --stream=xbstream --backup --target-dir /mnt/db/backup 2> /mnt/db/backup/stream_backup.log | pigz - > /mnt/db/backup/$(date +%F-%H-%M-%S).tar.gz
Most of the time we end up with this error message:
Error: failed to execute query ‘SELECT server_uuid, local, replication, storage_engines FROM performance_schema.log_status’: 1205 (HY000) Lock wait timeout exceeded; try restarting transaction
It also causes whole cluster to stop responding near the end of the backup (with “too many connections” error message), but resumes normal operations as soon as backup fails or completes succesfully.
We also tried the --safe-slave-backup but cluster lockup is even more noticable so we had to abort.