WAAS CancelableRequestTable Error

[corpitsol]

Hello,

We have one WAAS box producing this error continiously. I was not able to find any information to indicate what this "CancelableRequestTable" actually is and how it ties into CIFS.

Any ideas here?

2010-07-02 10:58:28,679  WARN (actona.cifs.smb.CancelableRequestTable:165) lightRxFlowPool-4 - Request ID: 2441/6989128 CancelableRequestTable has reached its limit (10)
2010-07-02 10:59:20,125  WARN (actona.cifs.smb.CancelableRequestTable:165) lightRxFlowPool-3 - Request ID: 2441/6996973 CancelableRequestTable has reached its limit (10)
2010-07-02 10:59:25,812  WARN (actona.cifs.smb.CancelableRequestTable:165) lightRxFlowPool-2 - Request ID: 2441/6997362 CancelableRequestTable has reached its limit (10)
2010-07-02 10:59:26,848  WARN (actona.cifs.smb.CancelableRequestTable:165) lightRxFlowPool-4 - Request ID: 2441/6997443 CancelableRequestTable has reached its limit (10)
2010-07-02 10:59:26,999  WARN (actona.cifs.smb.CancelableRequestTable:165) lightRxFlowPool-2 - Request ID: 2441/6997498 CancelableRequestTable has reached its limit (10)
2010-07-02 10:59:27,270  WARN (actona.cifs.smb.CancelableRequestTable:165) lightRxFlowPool-4 - Request ID: 2441/6997524 CancelableRequestTable

[Zach Seils · Accepted solution]

There is a limit (of 10) on the maximum number of outstanding change notifications per CIFS session.  This can happen when the client has multiple Explorer windows opened for different directories on the same file server.  If this happens and the file the change notification was issued for is changed directly on the file server, the client won't see the change immediately, but after ~30 seconds or a manual refresh.

Regards,

Zach

[corpitsol]

Zach,

This may be more of a training issue, but what is the best way to match/correlate these errors with a specific connection/user?

I wanted to shadow that endpoint and see if I could verify the behavior or multiple explorer windows or similar. The errors reference a transaction/session ID, but I don't have any easy way to search those against actual IP connections.

Matthew

[Zach Seils]

There are some additional debug messages that should help identify the user.  To change the log level of cifs_err.log, execute the following CLI command on the device where you see these log messages:

accelerator cifs expert attribute set Logger InternalLevel "DEBUG"

Wait a few minutes, then revert the setting with the following command:

accelerator cifs expert attribute set Logger InternalLevel "INFO"

Grab a new copy of the cifs_err.log file and post it here for review.

Thanks,

Zach

[corpitsol]

Attached: CIFS Logs, WAFS Benchmark, Userinfo (IP). I have a number of machines popping up with this problem.

Any insight would be appreciated. From the Connection status, it looks like the connection is being optimized by CIFS well, just tons of delays.

*The issue seems intermittent, and I havn't completely ruled out the machines themselves but I have some of the basics; All are new Windows 7, Dell laptops running in an SBS2008 fresh domain environment. Users will report sluggish mapped drive access and that the drives report unavailable, files are painfully slow to open and write etc... Hopefully we can accertain from these files if its a WAAS issue.

[Zach Seils]

I don't see any instances of the "CancelableRequestTable" error in the debug logs.  Is it possible to get the log over a longer period of time?  Are you trying to recreate the problem yourself using WAFS Benchmark, or are you capturing from one of the device where you see this message?

Thanks,

Zach