xtrabackup 2.2.3 failed: expected log block no. 435565628, but got no. 439759928 from the log file
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Percona XtraBackup moved to https://jira.percona.com/projects/PXB |
Fix Released
|
High
|
Sergei Glushchenko | ||
2.0 |
Won't Fix
|
Undecided
|
Unassigned | ||
2.1 |
Fix Released
|
High
|
Sergei Glushchenko | ||
2.2 |
Fix Released
|
High
|
Sergei Glushchenko |
Bug Description
Customer reported that xtrabackup 2.2.3 fails on taking incremental backups on one slave. Previous version of xtrabackup worked fine on this server before. And if the replication is stopped during the backup, backup would succeed.
Backup command:
-------
innobackupex --incremental --incremental-
Output:
----------------
140904 13:24:59 innobackupex: Starting ibbackup with command: xtrabackup_55 --defaults-
ed_2'
....
...
Variables:
-------
innodb_
innodb_
innodb_
innodb_
Status:
------------------
xtrabackup started at 13:24:59, failed at 13:58:06
At 13:28:05
Log sequence number 41454814261866
Log flushed up to 41454814255943
Last checkpoint at 41454812819846
At 13:54:04
Log sequence number 41456453356468
Log flushed up to 41456453355760
Last checkpoint at 41455591826645
Note:
--------------
1) This error looks similar to https:/
2) The two xtrabackup outputs from the customer(one above, the other one in comment #8 of bug #1073741823) showed the same number of different log blocks between the expected no. and got no., which is 4194300.
3) Xtrabackup output showed that the log sequence number started from 41454695642624 and never move beyond it.
Related branches
- Alexey Kopytov (community): Approve
-
Diff: 12 lines (+1/-1)1 file modifiedsrc/xtrabackup.cc (+1/-1)
- Alexey Kopytov (community): Approve
-
Diff: 12 lines (+1/-1)1 file modifiedstorage/innobase/xtrabackup/src/xtrabackup.cc (+1/-1)
tags: | added: i45208 |
Changed in percona-xtrabackup: | |
status: | Incomplete → New |
Changed in percona-xtrabackup: | |
assignee: | nobody → Sergei Glushchenko (sergei.glushchenko) |
Changed in percona-xtrabackup: | |
status: | Incomplete → New |
Please make sure that innodb_ log_file_ size used by the server and the value in my.cnf are the same.
One possible explanation for this bug could be that innodb_ log_file_ size value got changed in my.cnf without restarting the server. So InnoDB and PXB (which reads my.cnf) disagree on its value.