Views Notification Service Stalling/ Stopping

A

Angel

We have 3 project plans that seem to stall the views notification service
when it tries to process the XML files associated with these plans. They are
3 fairly large plans in terms of the number of assignments although not
really a large number of tasks. All three plans spanned the year of 2006 and
were recently closed out (all tasks marked 100% complete and remaining work
changed to 0, etc.)

However, everytime one of these plans is saved or published and the XML file
is generated, the Views Notification Service gets hung and must be stopped
(must use Task Manager to stop- it won't stop using the command options
within Services) and then restarted. I let the XML files sit around for
several hours to see if it was just really slow but it ended up never getting
processed, which in turn didn't allow our OLAP Cube to finish processing
since the XML file for it could never be processed.

I do not believe it's a hardware related issue as our system is quite
robust. We have 6 servers (DB server, front-end server, views notification
server, session manager, analysis server and a terminal server for folks who
work offsite). The Views Notification Server is running on a server with Dual
Xeon 3.6GHz processors and 4GB of RAM. The processor and memory usage are
normally barely used. When the server is trying to process an XML file
associated with one of these three plans, the CPU usage usually jumps up to
about 25% and but the memory usage is doesn't hardly change. I can watch the
VWNotify.exe service running in the task manager. It will run up to about
188kb for a minute or so and then drop down to somewhere in the 90s before it
freezes and doesn't change. This is usually the point where I stop the
service, remove the XML files from the ViewDrop folder and then restart it so
the other XML files can be processed.

We encountered this issue once last year but we never could narrow down the
issue to a single plan and eventually stopped encountering the issue.

We recently upgraded to SP2a and SP2 on all of our clients but we began
experiencing this issue approximately two weeks before (we were hopeful that
maybe the upgrade would resolve the issue). It has not.

The other abnormality with one of these three plans is that it contains
"false actuals". We use managed time periods and normally synching with
protected actuals will resolve any issue we may encounter. The synching
doesn't appear to actually be doing anything and no numbers are updated in
the plan. I have confirmed that synching does work in other plans so I
believe it's just a plan issue.

I read the post regarding the process of saving the file offline and then
importing it to force a rebuild of the binary record. Do you think this is an
instance where we might have corrupt data and need to do this? Are there any
known root causes of this problem other than abnormal characters in the title
or outline codes (there are none in any of these plans)? Any other
suggestions?

Thanks for your help.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top