[MDEV-20059] mariabackup-10.3.22 restore normal, but start service failed- InnoDB Trying to read page number 1858 in space 3, space name innodb_undo003, which is outside the tablespace bounds Created: 2019-07-13  Updated: 2020-03-02

Status: Open
Project: MariaDB Server
Component/s: mariabackup
Affects Version/s: 10.3.14
Fix Version/s: 10.3

Type: Bug Priority: Major
Reporter: sysdljr Assignee: Vladislav Lesin
Resolution: Unresolved Votes: 0
Labels: None
Environment:

Cent OS 7.3, MariaDB Server 10.3.14/10.3.18/10.3.22

Issue Links:
Relates
relates to MDEV-18194 Incremental prepare tries to access p... Closed

 Description   

Hi, all
Today, we use MariaDB full backup and increment backup file to restore:
– full restore
mariabackup --use-memory=8G --prepare --target-dir /data3/cluster04/data/

– incr restore
mariabackup --use-memory=8G --prepare --target-dir /data3/cluster04/data/ --incremental-dir /data3/incr/

both full restore and incr restore are normal, but start MariaDB failed,
error log:
2019-07-13 17:28:38 0 [ERROR] [FATAL] InnoDB: Trying to read page number 978 in space 3, space name innodb_undo003, which is outside the tablespace bounds. Byte offset 0, len 16384
190713 17:28:38 [ERROR] mysqld got signal 6 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
To report this bug, see https://mariadb.com/kb/en/reporting-bugs

We try to restore full and incrment backup again, error log show:
[ERROR] [FATAL] InnoDB: Trying to read page number 727 in space 2, space name innodb_undo002, which is outside the tablespace bounds. Byte offset 0, len 16384

after upgrade mariabackup version to 10.3.16, same error too.

if only restore full backup, MariaDB 10.3 server can start normal. MDEV-18194 fix similar question, Please help us to confirm , how to resolve the question ?

Thanks.

 Comments   
Comment by sysdljr [ 2020-03-01 ]

After upgrade to 10.3.18 , there six group instances restore and start normal in seven group instances,
but a instance restore normal, still start failed.

error log:
2020-03-01 12:59:54 0 [ERROR] [FATAL] InnoDB: Trying to read page number 1858 in space 3, space name innodb_undo003, which is outside the tablespace bounds. Byte offset 0, len 16384
200301 12:59:54 [ERROR] mysqld got signal 6 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
To report this bug, see https://mariadb.com/kb/en/reporting-bugs

after upgrade to 10.3.22 , still same error. please help confrim , thanks again.

in this instance , there a database about 1.2TB+ size, ten tables use page_compressed.

Generated at Thu Feb 08 08:56:24 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.