[MDEV-19305] Crash upgrade from previous 10.2 (minor upgrade) failed with "Cannot determine the space ID from the first 64 pages" Created: 2019-04-22  Updated: 2023-05-29  Resolved: 2023-05-29

Status: Closed
Project: MariaDB Server
Component/s: Storage Engine - InnoDB
Affects Version/s: 10.2
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Elena Stepanova Assignee: Elena Stepanova
Resolution: Won't Fix Votes: 0
Labels: None


 Description   

10.2 d315b4ff

 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Starting crash recovery from checkpoint LSN=49435537
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Checksum mismatch in datafile: ./test/t011.ibd, Space ID:33, Flags: 481
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size:1024. Pages to analyze:64
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size: 1024. Possible space_id count:0
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size:2048. Pages to analyze:64
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size: 2048. Possible space_id count:0
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size:4096. Pages to analyze:64
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size: 4096. Possible space_id count:0
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size:8192. Pages to analyze:64
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size: 8192. Possible space_id count:0
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size:16384. Pages to analyze:64
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size: 16384. Possible space_id count:0
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size:32768. Pages to analyze:32
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size: 32768. Possible space_id count:0
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size:65536. Pages to analyze:16
 
2019-04-23  0:50:14 139760450823616 [Note] InnoDB: Page size: 65536. Possible space_id count:0
 
2019-04-23  0:50:14 139760450823616 [ERROR] InnoDB: Datafile './test/t011.ibd' is corrupted. Cannot determine the space ID from the first 64 pages.
 
2019-04-23  0:50:15 139760450823616 [ERROR] InnoDB: Tablespace 33 was not found at ./test/t011.ibd.
 
2019-04-23  0:50:15 139760450823616 [ERROR] InnoDB: Set innodb_force_recovery=1 to ignore this and to permanently lose all changes to the tablespace.
 
2019-04-23  0:50:15 139760450823616 [ERROR] InnoDB: Plugin initialization aborted at srv0start.cc[2193] with error Tablespace not found
 
2019-04-23  0:50:15 139760450823616 [Note] InnoDB: Starting shutdown...
 
2019-04-23  0:50:15 139760450823616 [ERROR] Plugin 'InnoDB' init function returned error.
 
2019-04-23  0:50:15 139760450823616 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
 
2019-04-23  0:50:15 139760450823616 [Note] Plugin 'FEEDBACK' is disabled.
 
2019-04-23  0:50:15 139760450823616 [ERROR] Unknown/unsupported storage engine: InnoDB
 
2019-04-23  0:50:15 139760450823616 [ERROR] Aborting

To reproduce:

  • download and unpack the datadir ftp://perro.askmonty.org/public/jira/mdev19305.tar.gz
  • start 10.2 server with 

    --file-key-management --file-key-management-filename=`pwd`/mysql-test/std_data/keys.txt --plugin-load-add=file_key_management.so --innodb-encrypt-tables --innodb-encrypt-log --innodb-encryption-threads=4 --aria-encrypt-tables=1 --encrypt-tmp-disk-tables=1  --innodb-page-size=64K --innodb-compression-algorithm=none --loose-innodb-file-format=Barracuda --loose-innodb-file-per-table=1  --loose-innodb_log_compressed_pages=on

    (check if this form of file-key-management-filename is applicable in your case).

The data directory was created using 10.2.23 release server, which was killed by SIGKILL during execution of DML. Server error and general (SQL) logs can be found in the same datadir.
No obvious signs of hardware failure were observed during the test.

 Comments   
Comment by Marko Mäkelä [ 2023-05-29 ]

I’m terribly sorry for having missed this.

In MariaDB Server 10.6, the recovery of DDL operations was rewritten e.g., in MDEV-24626 and MDEV-25506, to strictly follow a write-ahead-logging protocol. Are this kind of errors reproducible on 10.6 or later?

Comment by Elena Stepanova [ 2023-05-29 ]

It's a 4-year old bug filed against an already EOL-ed version. I see no point trying to handle it now.

Generated at Thu Feb 08 08:50:38 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.