Bug #21270 database corrupted
Submitted: 25 Jul 2006 10:58 Modified: 25 Aug 2006 17:38
Reporter: sinead casey Email Updates:
Status: No Feedback Impact on me:
None 
Category:MySQL Server: InnoDB storage engine Severity:S2 (Serious)
Version:4.1 OS:Windows (Windows 2003 Server)
Assigned to: Assigned Account CPU Architecture:Any
Tags: database corrupt

[25 Jul 2006 10:58] sinead casey
Description:
After mysql closed down unexpectedly we received the following error:

060718 11:02:59  InnoDB: Error: page 1681947 log sequence number 38 934260932
C! Current system log sequence number 38 870148736.
InnoDB: Your database may be corrupt.

While this error was occuring we were still able to write to the database. Subsequently we received the following error and no writes or reads were possible:

060718 22:18:34  InnoDB: error clustered record for sec rec not found
InnoDB: index `FK819EA42F5A3F78ED` of table `mscapelivejan252006/t_hourly_fact`
InnoDB: sec index record PHYSICAL RECORD: n_fields 8; 1-byte offs TRUE; info bits 0
 0: len 4; hex 800006a5; asc    ¥;; 1: len 8; hex 800000000000000f; asc         ;; 2: len 4; hex 80000005; asc     ;; 3: len 4; hex 8000002e; asc    .;; 4: len 4; hex 80000004; asc     ;; 5: len 4; hex 80000001; asc     ;; 6: len 4; hex 80000001; asc     ;; 7: len 8; hex 800000000004e2fb; asc       âû;;

InnoDB: clust index record PHYSICAL RECORD: n_fields 16; 1-byte offs TRUE; info bits 0
 0: len 8; hex 800000000000000f; asc         ;; 1: len 4; hex 80000005; asc     ;; 2: len 4; hex 8000002e; asc    .;; 3: len 4; hex 800006a5; asc    ¥;; 4: len 4; hex 80000004; asc     ;; 5: len 4; hex 80000001; asc     ;; 6: len 4; hex 80000001; asc     ;; 7: len 8; hex 800000000004e266; asc       âf;; 8: len 6; hex 000025f3c1cd; asc   %óÁÍ;; 9: len 7; hex 80001040020084; asc    @  „;; 10: len 8; hex 8000000000000001; asc         ;; 11: len 8; hex 8000000000002308; asc       # ;; 12: len 8; hex 8000000000002308; asc       # ;; 13: len 8; hex 8000000000002308; asc       # ;; 14: len 8; hex 8000000000000000; asc         ;; 15: len 8; hex 8000000000000000; asc         ;;

TRANSACTION 0 656927543, ACTIVE 1086 sec, OS thread id 7964 fetching rows, thread declared inside InnoDB 308
mysql tables in use 3, locked 0
MySQL thread id 18, query id 1096950 localhost 127.0.0.1 root Sending data
SELECT AL3.c_category, AL3.c_name, innr.volume, innr.total_resp, innr.max_resp, innr.faults FROM (SELECT AL1.c_id_service, SUM(AL1.c_metrics_number_of_events) as volume, SUM(AL1.c_metrics_total_response_time) as total_resp, MAX(AL1.c_metrics_max_response_time) as max_resp, SUM(AL1.c_metrics_number_f

InnoDB: Submit a detailed bug report to http://bugs.mysql.com
060718 22:19:52  InnoDB: error clustered record for sec rec not found
InnoDB: index `FK819EA42F5A3F78ED` of table `mscapelivejan252006/t_hourly_fact`
InnoDB: sec index record PHYSICAL RECORD: n_fields 8; 1-byte offs TRUE; info bits 0
 0: len 4; hex 80000745; asc    E;; 1: len 8; hex 800000000000000f; asc         ;; 2: len 4; hex 80000018; asc     ;; 3: len 4; hex 80000005; asc     ;; 4: len 4; hex 80000004; asc     ;; 5: len 4; hex 80000001; asc     ;; 6: len 4; hex 80000001; asc     ;; 7: len 8; hex 800000000004e2f5; asc       âõ;;

Would you be able to shed any light as to what may have happened here. The dtabase was quite large at the time 40 GIG. Thanks

How to repeat:
Error was consistent once it started may if been a consequence of the first error and the size of the database (40 GIG) 
 
060718 11:02:59  InnoDB: Error: page 1681947 log sequence number 38 934260932
C! Current system log sequence number 38 870148736.
InnoDB: Your database may be corrupt.

Suggested fix:
We needed to create a new database, appears the database was corrupted.
[25 Jul 2006 17:38] Valeriy Kravchuk
Thank you for a problem report. Please, specify the exact (4.1.x) version used. Any ideas on how to repeat the behaviour described are also welcomed.

Are you sure it is not a result of hardware failure?
[27 Jul 2006 5:14] Heikki Tuuri
Sinead,

please post the ENTIRE .err log from the server. Do not cut anything off.

A possible reason for the corruption is that your server failed to write ib_logfiles to disk. If you have manuals of the disk system, please check if you have write caching enabled in the disk system even though the cache is not battery-backed.

You may be able to save most of your data by dumping and reloading the corrupt tables.

Regards,

Heikki
[25 Aug 2006 23:00] Bugs System
No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".