NBU 4.5 Status 150 - scheduler hang problem

[Estoragetechie]

Hi, I'm running NBU4.5 MP8 on Windows 2003 - usual parameters all in place, tuned for Gigabit ethernet. However we have been having regular (weekly) occurrences of all backups fialing with 150. Tracked it down to the scheduler log file:

21:03:41.913 [5288.2752] <2> send_runQ_msg_list: Deferring due to message write failure. errno=11
21:03:41.913 [5288.2752] <2> send_runQ_msg_list: Deferring due to message write failure. errno=11
21:03:41.913 [5288.2752] <2> send_runQ_msg_list: Deferring due to message write failure. errno=11
21:03:41.913 [5288.2752] <2> send_runQ_msg_list: main_sched suspending---possible hang condition
21:03:41.913 [5288.2752] <2> send_runQ_msg_list: queue full, suspending process 1
21:03:52.632 [5288.2752] <2> send_runQ_msg_list: unsuspending process 1

and then the backup works - but it happens for all jobs!

Obviouslyt the message queue is full but how can I tweak this in windows? Any ideas?

(server is IBM maxed out on memory - just installed up to 8GB)

 

[Stumpr]

Windows tuning:

http://www.zedat.fu-berlin.de/services/fab/man/4.5/TuningNetbackupOnNT-4.5.pdf

Bob Stump
Just because the Veritas documentation states a certain thing does not make it a fact

 

[Estoragetechie]

Thanks for the response Bob - I have been through that. I am more inclined to think we have an OS issue or a bpsched issue.

If this was unix I'd be checking all the share memory /semaphore message queues settings and DNS response issues - actually that might be an idea. Seen the message queues fill up before due to DNS flooding (caused by reverse name lookups). Logged a call with Veritas via 3rd party so watch this space...

Al.