Network flooding on DL380

[Joeri Boeree]

Hi,

I'm currenlty facing some big troubles with a DL380 G3 server which connect to a Procurve 6108. The problems which occur are wsimilar to this MS article: http://support.microsoft.com/?kbid=317249

The server performance breaks down and event id's 2021 and 2022 are logged. When we disconnect the cable the server immediatly pops up again and you are able to work normally again. then we conenct the cables and it works for one hour to a day and than the problems occur again.

Main errors in the event log are id 2021 and 2022 which refer to network load difficulties.

All the given solutions in the MS-doc are allready tried. Many other possible solutions are also tried like fixing speed (AUTO1000, 100 full), auto speed on both server and switch.

Right now one of the servers has been connected to a 2524 at 100 full duplex just to see how this works.

All firmwares are up to date (7.32 for 6108), all drivers are Smartstart 6.40.

Can this be a compatibility problem? Please help.

Thanks, Joeri.

[Mark Medici]

Sounds like you need to find out where the network traffic is coming from. Mirror the server's switch port to another port and use a sniffer. If you don't have a sniffer, you can try the free Ethereal sniffer from http://www.ethereal.com/.

[John Marinov]

How much memory is in the server, and is it swapping.

We have seen 2012 and 2022 errors disappear after increasing memory.

Are you using teaming?

Regards

[Matt Shive]

Hello,
Curious if you found a solution.

I am having a very similar problem with a BL20p blade. 2-node Windows Server 2003 fileserver cluster. Network timeouts come and go and the only clues are the event 2021 warnings and an occasional event 2020 error from the srv source.

Any help or advice is appreciated.

Matt Shive
Intel Technology Manager - EMEA
Black & Decker, Inc.