MySQL Bugs: #49748: WIndows InnoDB: Operating system error number 1784 in a file operation

Bug #49748	WIndows InnoDB: Operating system error number 1784 in a file operation
Submitted:	16 Dec 2009 21:11	Modified:	31 Aug 2010 7:22
Reporter:	James Day	Email Updates:
Status:	Can't repeat	Impact on me:	None
Category:	MySQL Server: InnoDB storage engine	Severity:	S2 (Serious)
Version:	5.0.18, 5.0.51b, 5.1.47	OS:	Windows
Assigned to:		CPU Architecture:	Any

Description:
There are occasional reports of "Operating system error number 1784 in a file operation" in the InnoDB error log followed by a shutdown. One example is from a comment on bug #3139:

 ----

[30 Nov 19:45] sadasasd sdasd 
Not quite sure if it's same issue but we are having a very similar error and I was wondering if this patch fixes this problem as well:

InnoDB: Operating system error number 1784 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 D:\MySQLlog\ib_logfile0
InnoDB: File operation call: 'Windows aio'.
InnoDB: Cannot continue operation.
InnoDB: Log scan progressed past the checkpoint lsn 35 1443981126   

[30 Nov 19:59] Shane Bester 
error 1784 means ERROR_INVALID_USER_BUFFER (The supplied user buffer is not valid for the requested operation.)

 ----

And from the MySQL lists:

http://lists.mysql.com/mysql/219544

 ----

Another from the forums:

http://forums.mysql.com/read.php?22,260318,260318

 ----

How to repeat:
Unknown. Possibly related to bug #48946.

Suggested fix:
A possible workaround from the lists example:

"The problem is solved I guess, no problems reported after I made the changes. As suggested, increased the log file sizes.

Changed innodb_buffer_pool_size from 128M to 256M.
Changed innodb_log_file_size from 5M to 64M.

Also changed innodb_thread_concurrency from 20 to 8. This might have helped
as well."

> # ERROR_INVALID_USER_BUFFER or ERROR_NOT_ENOUGH_MEMORY Each device
>  driver maintains a fixed-size list (in a nonpaged pool) of
>  outstanding I/O requests. If this list is full, the system can’t
>  queue your request, ReadFile and WriteFile return FALSE, and
>  GetLast-Error reports one of these two error codes (depending on the
>  driver).

Possibly caused by a shortage of non-paged pool memory when the InnoDB log file size is too small. Presumably retrying in the case of this error number would make sense

I think it would make more sense to start upgrading the server to 5.0.88 (if you are serious that you are using 5.0.18) and verify if problem persists.

Peter
(not a MySQL person)

Peter, thanks. That's just the only version that was given in the reports of the problem. With two reports as recent as the last month it's unlikely that it's just due to old versions being used.

James Day, MySQL Senior Support Engineer, Sun Microsystems

ERROR_INVALID_USER_BUFFER sometimes really means "invalid user buffer", that is read buffer in ReadFile does not point to valid memory location

We had a similar issue recently ourselves. It ended up being a corrupted .ini (configuration) file (not my.ini, but our SQLyog.ini was corrupted on this users' system).  The file length as reported by the file system was not correct as compared to the actual file content. 

Not sure about NTFS details, but NTFS stores a lot of metadata in a database much similar to the Windows registry.  And something was wrong here it seems.

Peter,

do you have any idea how to repeat this situation?

> . It ended up being a corrupted .ini
(configuration) file (not my.ini, but our SQLyog.ini was corrupted on this users'
system).  The file length as reported by the file system was not correct as compared to
the actual file content. 

Can this mean you have problems with hard disk?

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".

Can innodb retry the failed operation a few times in case it works ?

http://msdn.microsoft.com/en-us/library/cc500408.aspx

"If GetLastError returns a value other than ERROR_IO_PENDING, the I/O request could not be queued to the device driver. Here are the most common error codes returned from GetLastError when an I/O request can’t be queued to the device driver:

ERROR_INVALID_USER_BUFFER or ERROR_NOT_ENOUGH_MEMORY Each device driver maintains a fixed-size list (in a nonpaged pool) of outstanding I/O requests. If this list is full, the system can’t queue your request, ReadFile and WriteFile return FALSE, and GetLast-Error reports one of these two error codes (depending on the driver)."

I am sorry for overlooking/forgetting Sveta's question to me.  But this report is now so old that I cannot memorize and probably not reconstruct environment.

But Shane seems to be on track here.