Bug #20827 Bug 3139 closed, but no resolution???
Submitted: 3 Jul 2006 16:51 Modified: 11 Nov 2008 6:47
Reporter: Ken Hanks Email Updates:
Status: No Feedback Impact on me:
None 
Category:MySQL Server Severity:S1 (Critical)
Version:4.1.15, 5.0.36BK OS:Windows (Windows Server 2003, WinXP 64-bit)
Assigned to: CPU Architecture:Any

[3 Jul 2006 16:51] Ken Hanks
Description:
MySQL: ready for connections.
Version: '4.1.15'  socket: ''  port: 3306  Official MySQL binary
060620 16:11:08  InnoDB: Operating system error number 995 in a file operation.
InnoDB: Some operating system error numbers are described at
InnoDB: http://dev.mysql.com/doc/mysql/en/Operating_System_error_codes.html
InnoDB: File name D:\mysql\data\ibdata1
InnoDB: File operation call: 'Windows aio'.
InnoDB: Cannot continue operation.

I read the previous bug report (3139) but I don't see a resolution or a workaround except that it "MIGHT" be an OS bug. So I don't understand why the previous bug was closed.

This happens to us several times a day. Yes, we have a couple of large (10 GB) tables but they only contain 2 columns -- an id and a blob that contains scanned images, pdf files, etc.. , so it's not like I can normalize them or denormalize them.

Am I going to have to go with that database from the Evil Empire to get around this??

How to repeat:
Use large INNODB tables. Use Windows (we use server 2003 but I'm not sure that it doesn't occur on other versions). Work 8 hours. It'll happen.
[4 Jul 2006 8:32] Valeriy Kravchuk
Thank you for a problem report. It was decided that 3139 is caused by Windows bug, not a bug in MySQL code.

Anyway, if you have a repeatable exact sequence of actions that leads to this bug (on MySQL 4.1.20) each and every time, please, send it. Many sites use hundreds of GB of data in InnoDB tables on Windows without any problems of this kind.
[4 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".
[20 Feb 2007 7:29] MySQL Verification Team
During stress tests on a 64-bit build of 5.0.36 I received the error on home PC.
Version: '5.0.36'  socket: ''  port: 3306  yes
070220  3:55:16  InnoDB: Operating system error number 995 in a file operation.
InnoDB: Some operating system error numbers are described at
InnoDB: http://dev.mysql.com/doc/refman/5.0/en/operating-system-error-codes.html
InnoDB: File name .\ibdata1
InnoDB: File operation call: 'Windows aio'.
InnoDB: Cannot continue operation.

Disk is local SATA drive.  At the time of the error, 30 threads ran inserting/deleting/updating 1 table.  Also, a CHECK TABLE was running.
The client app (on same machine) was reporting some out of memory errors around the time of the error message.  See complete attached log.
[20 Feb 2007 7:46] MySQL Verification Team
some info about when error occurred.

Attachment: bug20827.log (application/octet-stream, text), 143.00 KiB.

[3 May 2007 14:18] Stephan Eggy
I am having the same errors, the mysql service crashes several times a day. I use windows server 2003 with plesk.

This are the errors taht are displaying:
070503 10:10:07  InnoDB: Operating system error number 995 in a file operation.InnoDB: Some operating system error numbers are described atInnoDB: http://dev.mysql.com/doc/refman/5.0/en/operating-system-error-codes.htmlInnoDB: File name .\ibdata1InnoDB: File operation call: 'Windows aio'.InnoDB: Cannot continue operation.

I will attach my error log to the files.
[3 May 2007 14:23] Stephan Eggy
This is the error file

Attachment: error.log (application/octet-stream, text), 128.15 KiB.

[3 May 2007 15:17] MySQL Verification Team
Hi Stephan,

Thanks for the logs, please check your OS isn't running out of memory, or swapping.
Also, can you monitor the disk i/o queued operations using perfmon?

There is a counter called 'current disk queue length' in the LogicalDisk and PhysicalDisk section that might shed more light.  What disk subsystem do you have? Specifically, SAN or local, and drivers, etc.
[3 Jun 2007 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".
[10 Oct 2008 22:33] James Day
MSDN : error 995 The I/O operation has been aborted because of either a 
thread exit or an application request. ERROR_OPERATION_ABORTED 

The thread probably had a failed memory allocation request or was being killed. Check that the server isn't allocating too much RAM.
[12 Nov 2008 0:00] 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".