Backup to gets stuck after non-InnoDB tables

MySQL & MariaDB Percona XtraBackup
1

We’re having a few issues with xtrabackup 8.0.14 on mysql 8.0.19.

The backup runs fine at first but then gets stuck at log scanned up to (some number) and the backup directory stops gaining size on the disk.
Backup gets started using the following command

xtrabackup --user=root --password=<snipped> --backup --target-dir /mnt/backup --compress --throttle 50

Executing FLUSH NO_WRITE_TO_BINLOG BINARY LOGS                                                                                                                                                       
Selecting LSN and binary log position from p_s.log_status                                                                                                                                            
Executing FLUSH NO_WRITE_TO_BINLOG ENGINE LOGS...                                                                                                                                                    
xtrabackup: The latest check point (for incremental): '605109943215'                                                                                                                                                 
xtrabackup: Stopping log copying thread at LSN 606605028003.                                                                                                                                                         
log scanned up to (604684959232)                                                                                                                                                                  
log scanned up to (604684959232)                                                                                                                                                                  
log scanned up to (604684959232)                                                                                                                                                                  
log scanned up to (604684959232)                                                                                                                                                                  
log scanned up to (604684959232)

These are the last lines printed with informations. After that, endless lines of log scanned up to…

2 Likes
2

Is there any SHOW PROCESS activity for the backup user?

2 Likes
3

Yes. There is a sleeping connection up to the time of the mentioned log lines. Then obviously one query gets executed and the sleeps time resets to zero.
After that nothing happens. Just the sleep timer counting up

2 Likes
4

Hi @Patrick_Sehorz . The step your backup is at it means it’s waiting the redo log copy thread to catch-up until it copies the LSN 606605028003. However, we can see that the log copying thread is stuck on LSN 604684959232 . The difference between were it has to stop and where it is at the moment is around 1.78G .
I would like to check your my.cnf to validate if you have that space available on redo logs, to rule out a bug where the logs have wrapped around and the tool has not identified it, since you have throttled the io on your backup.

Does it happens all the time or it was a one time issue?

2 Likes
5

Closing due to inactivity