FSM crashing Windows 2000 Server - Event IDs 12042 and 12052

[Noah Lee]

I've looked everywhere but I cannot find an answer. We have an external IDE
RAID system that we have hooked up to a Windows 2000 Server via SCSI. We
shared this with File Services for Macintosh and things seemed fine as long
as our clients were using OS9.x. They all recently upgraded to OSX and as
soon as they did and accessed this macintosh share the server would reboot.

We have a bunch of errors in the Event Log, listed as event id 12042 and a
few for 12052:

Event ID: 12042
Source: Macfile
Type: Error
Internal server information for file
"\Projects\LoW3\Backup\Entrances\Build\Xbox\0002683.mdf:AFP_AfpInfo" was
corrupted. Setting default information.

Event ID: 120052
Source: Macfile
Type: Error
A directory change notification was missed on volume "Projects".

Any ideas what would cause these? We've tried two different Windows 2000
Servers, with both SP3 and SP4, both with the same results. We have another
one of these Promise IDE RAID systems hooked up to a different Windows 2000
server with no problems. The 2 folders we shared have a combined 955,651
files and 7084 folders. Is there a limitation there that would cause this?
The server that the RAID is currently on has 1GB of memory.

Any help appreciated!

 

[William M. Smith]

I've looked everywhere but I cannot find an answer. We have an external IDE
RAID system that we have hooked up to a Windows 2000 Server via SCSI. We
shared this with File Services for Macintosh and things seemed fine as long
as our clients were using OS9.x. They all recently upgraded to OSX and as
soon as they did and accessed this macintosh share the server would reboot.

We have a bunch of errors in the Event Log, listed as event id 12042 and a
few for 12052:

Event ID: 12042
Source: Macfile
Type: Error
Internal server information for file
"\Projects\LoW3\Backup\Entrances\Build\Xbox\0002683.mdf:AFP_AfpInfo" was
corrupted. Setting default information.

Event ID: 120052
Source: Macfile
Type: Error
A directory change notification was missed on volume "Projects".

Any ideas what would cause these? We've tried two different Windows 2000
Servers, with both SP3 and SP4, both with the same results. We have another
one of these Promise IDE RAID systems hooked up to a different Windows 2000
server with no problems. The 2 folders we shared have a combined 955,651
files and 7084 folders. Is there a limitation there that would cause this?
The server that the RAID is currently on has 1GB of memory.

Hi Noah!

I don't have anything definitive for you, but this sounds like you might
have some failing hardware. When you tested the two servers, were you using
the same RAID hardware and drives?

Can you double-check the number for the 12052 Event ID? My reference says
the source should be a POP3 Connector. I couldn't find a 120052 reference.

You're no where near the file count limits for Mac File Services and you
have plenty of memory.

If you haven't already, you might also try deleting and then recreating the
Mac volume in case it may have become corrupt.

bill

 

[Noah Lee]

Hi Noah!

I don't have anything definitive for you, but this sounds like you might
have some failing hardware. When you tested the two servers, were you using
the same RAID hardware and drives?

Can you double-check the number for the 12052 Event ID? My reference says
the source should be a POP3 Connector. I couldn't find a 120052 reference.

You're no where near the file count limits for Mac File Services and you
have plenty of memory.

If you haven't already, you might also try deleting and then recreating the
Mac volume in case it may have become corrupt.

Yes I'm using the same Promise RAID tower then recreating the shares on
them, on two different Windows 2000 servers. As soon as they're done
indexing and the people start accessing the share the server will start
randomly rebooting.

Maybe it is hardware??

The errors are 12042 and 12052, which is weird because I cannot find any
reference to those event id's *anywhere*.

 

[Noah Lee]

I talked to my boss and was reminded that this problem has actually followed
this set of data around in that when it was stored on the actual server (not
the promise IDE RAID tower), it would have the same symptoms. We ended up
purchasing Extreme Z-IP from GroupLogic to help combat this, but we'd like
to just use File Sharing for Mac like we are on our other servers...

Maybe that info helps?

 

[William M. Smith]

I talked to my boss and was reminded that this problem has actually followed
this set of data around in that when it was stored on the actual server (not
the promise IDE RAID tower), it would have the same symptoms. We ended up
purchasing Extreme Z-IP from GroupLogic to help combat this, but we'd like
to just use File Sharing for Mac like we are on our other servers...

Hi Noah!

I'll plug http://www.eventid.net/ for referencing ID numbers. It's very
handy.

12042 is Mac-related and according to EventID.net the problem may lie in a
file name or file names.

If this Mac volume was created after having been populated with files, then
the files on it may contain invalid characters or may be longer than 31
characters, which is a limit of the Apple Filing Protocol 2.2, which Windows
2000 servers use.

Macs can handle just about any character that Windows can use although a
colon ":" is out of the question because it's a directory character much
like a slash is in Windows.

I would guess your problem may lie with file names that are too long and
when someone tries to access one of these files, AFP is crashing your
server. The fact that you mention Extreme Z-IP almost reinforces this since
this product uses AFP 3.x, which does support long filenames up to 255
characters.

Hope this helps! bill