MySQL Bugs: #28120: Service cannot be started again after restart!!

Bug #28120	Service cannot be started again after restart!!
Submitted:	26 Apr 2007 13:06	Modified:	12 Mar 2008 14:03
Reporter:	Samoil Bugediwala	Email Updates:
Status:	Duplicate	Impact on me:	None
Category:	MySQL Server	Severity:	S1 (Critical)
Version:	5.0.45, 5.0.37	OS:	Windows (XP Professional OEM)
Assigned to:		CPU Architecture:	Any

Description:
We are trying to install MySQL 5.0.18 on a Dell Optiplex Gx745 with RAID 1 configuration. The installation is perfect and no problems with it.

Problem:
Everything works fine till u restart the service or the PC as the service cannot be restarted. After restarting the PC if I try to connect using MySqlAdministrator it gives the following error.

Could not connect to the specified instance.

MySQL Error Number 2003
Can't connect to MySQL Server on 'localhost'(10061)

I am using an Adaptec Raid Card for RAID1 configuration.

--------------------------------------------------------------------

Below is the error log:

070426 5:25:24 [Note] C:\Program Files\MySQL\MySQL Server 5.0\bin\mysqld-nt: Normal shutdown

070426 5:25:25 InnoDB: Starting shutdown...
070426 5:25:27 InnoDB: Shutdown completed; log sequence number 0 43655
070426 5:25:27 [Note] C:\Program Files\MySQL\MySQL Server 5.0\bin\mysqld-nt: Shutdown complete

070426 5:26:34 InnoDB: Operating system error number 87 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 .\ib_logfile0
InnoDB: File operation call: 'aio read'.
InnoDB: Cannot continue operation.
070426 5:35:13 InnoDB: Operating system error number 87 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 .\ib_logfile0
InnoDB: File operation call: 'aio read'.
InnoDB: Cannot continue operation.

---------------------------------------------------------------------------
I believe this is a problem with RAID because when we use MySQL on the same machine without RAID it works fine.

Can someone please help me to solve the problem as this is a part of a big project and we are all stuck at this point.

How to repeat:
Installing MySQL 5.0.xxx on a RAID1 configuration should help in replicating the problem.

Let me know if more details are required.

Thank you for a problem report. Please, try to repeat with a newer version, 5.0.37. In case of the same problem, please, check if adding

innodb_flush_method=unbuffered

to my.ini makes any difference. Are you sure your RAID-1 working properly? 

(Read bug #8021 and bug #3139 for similar problems)

Thank you for the reply.

I have already tested this MySQL v5.0.37 and I have also checked with alpha & beta versions. But still the problem is the same.

I have doubled checked RAID1 & also have tried re-installing RAID but still the same problem remains with MySQL.

Thank you,
Samoil

Thank you for a dditional checks. Have you checked with innodb_flush_method=unbuffered?

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".

Same problem here, RAID 1 setup. I added innodb_flush_method=unbuffered to the default setup and it still will not restart the service.

Using MySql 5.0.45 for Windows 32 bit
I have an Adaptec Serial ATA Raid Card in the machine.

------ Error Log ----------------------------------------------------

080117 15:56:07  InnoDB: Starting shutdown...
080117 15:56:09  InnoDB: Shutdown completed; log sequence number 0 43655
080117 15:56:09 [Note] C:\MySql\bin\mysqld-nt: Shutdown complete

080117 15:56:10  InnoDB: Operating system error number 87 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 .\ib_logfile0
InnoDB: File operation call: 'aio read'.
InnoDB: Cannot continue operation.
080117 15:56:39  InnoDB: Operating system error number 87 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 .\ib_logfile0
InnoDB: File operation call: 'aio read'.
InnoDB: Cannot continue operation.
080117 15:56:57  InnoDB: Operating system error number 87 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 .\ib_logfile0
InnoDB: File operation call: 'aio read'.
InnoDB: Cannot continue operation.

I still think this is drivers-related problem. Please, upgrade XP and RAID drivers to latest versions available and inform about any results.

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".

I have the same problem. I cannot start mysql as a service (the same problem: Operating system error number 87). 
It starts only first time after instaling mysql - then mysql creates files: ibdata1, ib_logfile0, ib_logfile1 and works perfectly. After restarting service it cannot run.

Instead of that error, I can start mysql manualy by command line:

mysqld-nt --defaults-file="C:\[...path...]\my.ini"

Then it works witout any problem. 

My mysql version is 5.0.37, Windows XP SP2, my mainboard: Supermicro PDSME+, Xeon 3050, 2gb ram, RAID 1.

After upgrading to 5.0.51a I cannot start mysql (even manualy), so I can use only 5.0.37.

Right now I found, that there is the same problem that is in bug #28913. If I put line in my.ini:

innodb_flush_method=normal

mysql works perfectly.

I'll set this as duplicate of bug #28913 because there is a better description at this bug.

The problem might be a file system block size that is > 16 kB. See the other bug report.