[2007] Ready for processing - Waiting for process... for days

R

Robert X

Sorry for the cross post. I was inadvertently on the Developer ng :-(

Hello,

The topic of a lot of posts in this news group is "how to kill the queuing
process" which locks the Check In process, which locks the Save process...
My question is: did you notice a specific set of circumstances which
generates these Queuing locks. How to avoid them?
NB: I really miss the spooler errors of PS 2003: they where so gentle!

Robert
 
M

Marcin Chomicz

Robert
I ecountered this problem few times.
My test platform is single server based,on virtual platform.
I did not find any particular reason of them, but,
- I have reconfigured SQL connections enabling TCP/IP and NamedPipes
- I assigned more memory to the server (1024 now)
for about 3 days I haven't seen this problem (manipulating from three
clients in paralel, both PWA and PrjPro)
I can't find any idea why the problem has dissapeared.
best
 
R

Robert X

Hello Marcin,
Thanks for sharing your experience.
My test platform is single server based too. But i don't use any virtual
machine.
I have 2 Go RAM.

Robert
 
R

Rod Gill

Is SQL Server on the same box? If so, then often the Queuing service can
start before SQL Server gets up and running so it can't read/write to the
job Queue in SQL Server. Solution is to either move SQL Server to a
different box, or restart the Queuing service after a reboot and all other
services have started.

--

Rod Gill
Project MVP

Project VBA Book, for details visit:
http://www.projectvbabook.com

NEW!! Web based VBA training course delivered by me. For details visit:
http://projectservertraining.com/learning/index.aspx
 
R

Robert X

Rod,

My SQL Server 2000 is indeed on the same box than Project Server 2007.
So I notice that for a production environment, it's better to install SQL
Server on a dedicated server.
But on my single server, the 3 services : sqlmangr.exe, sqlagent.exe,
sqlservr.exe are already running.
How can the Queuing service start before SQL Server gets up and running ?
Thanks
 
S

Sharry Heberer [MSFT]

Actually, the issue is usually that SharePoint is somehow not ready yet
(possibly due to timing issues with SQL on startup also), when the Queue
asks SharePoint what PWA sites it needs to service, so no sites are
returned, and therefore the Queue thinks it has nothing to monitor and the
proper processes do not spawn. This is why you will still see the service
process (the Queuing.exe process running under NetworkService account) and
no others. The proper number of Queuing.exe processes that should be
running for your Queue to operate correctly is (# of SSPs + 1).

Hope this helps,
Sharry
 
D

Dale Howard [MVP]

Sharry --

When I have run into this issue, I followed the steps recommended in the
early PS2007 beta days, which were:

1. Stop the WSS Timer Service
2. Stop the Queueing Service
3. Start the Queueing Service again
4. Start the WSS Timer Service again

Is this what you would still recommend? Let us know. Thanks!
 
S

Sharry Heberer [MSFT]

Absolutely not. In fact, as my post mentions, if the Queue is processing
other jobs just fine, and it's only a few projects that are stuck, or if the
Queue is processing *any* jobs just fine (like maybe some Reporting jobs or
Cube Building jobs or whatever), then restarting the Queue won't get you
anything but unknown data loss. For instance, if there are jobs currently
being sent in via someone doing a Project Save via WinProj, and you restart
the Queue, that job will not finish being sent in, and you will have data
loss (whatever data was being sent from WinProj). This is because when the
Queue comes back up, it knows it did not receive the entire job, and so it
will ignore that "getting enqueued" job and cancel it. But an admin should
really be doing this step manually, so that they know what projects are
having problems, and they can notify the sender or take other appropriate
action.

When you follow the steps in my last post, you get to know *exactly* what
you are affecting by taking whatever actions you take. This is the proper
way an admin should be interacting with the Queue.

Since restarting the Queue will only help if *no* jobs are processing, it is
only recommended if you do find that *no* jobs are processing at all. To
test your Queue processing out, try to submit a new job, like create a fake
resource or something, and see if the Reporting Resource Sync goes through
the Queue successfully. If it does, then restarting the Queue won't help
(or even if it does *appear* to help, it only means that there was a job
stuck in "getting enqueued" state and restarting the Queue just wiped that
job out).

Gosh, I'd like to know who gave out the instructions to restart the Queue if
you have a problem. That is totally the wrong approach.

--
This posting is provided "AS IS" with no warranties, and confers no rights.

Use of included script samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm
 
S

Sharry Heberer [MSFT]

Oh my goodness. This thread got splintered and I forgot about that. I
totally didn't even answer the question you were asking! OK, let me answer
your real question...

So I think you're asking what the proper way to restart the Queue is when
you have SQL and PS all on one machine, and the Queue doesn't know there are
any sites to service because of the timing issue on startup. If so, then
yes - the steps you have for restarting the Timer service and the Queue
service are correct.

SO sorry for my mistake! I wish there was some kind of recall I could do on
this last post! :)

--
This posting is provided "AS IS" with no warranties, and confers no rights.

Use of included script samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm
 
D

Dale Howard [MVP]

Sharry --

Thanks for your sobering words! :)

If found that recommendation in the beta site used for reporting bugs in
Project Server 2007. It was a Microsoft staff member who proposed the fix,
but it was only for getting the Queue unstuck when it was not processing ANY
jobs.
 
D

Dale Howard [MVP]

Sharry --

Your last post was sage, even if incomplete. Since I use the virtual image
of Project Server 2007 from Microsoft, I occasionally need to restart the
Queue when it gets stuck. The steps always seem to help, but I was
wondering if they were the correct steps. Apparently the answer is yes for
my particular situaton. Thanks for sharing! :)
 
S

Sharry Heberer [MSFT]

OK, so maybe I did get your question right the first time. Oh well, I
answered multiple things then. :)

Yes - correct. If the Queue is not processing *any* jobs (or, more
precisely, if you notice that your Timesheet Queue is processing fine, but
your Project Queue is not, or vice versa), then your only recourse is to
restart the WSS Timer/Queue services as you have outlined (caveat: this
assumes you have already checked the obvious, like network conn, DB server
is up, etc.).

The most important thing to be aware of is that restarting the Queue means
that all SSPs are affected and will briefly lose servicing by the Queue
during the restart. This is why restarting is never recommended except as a
last resort. You just don't know what data is being sent in to what PWA
site at the time you hit that restart button.

--
This posting is provided "AS IS" with no warranties, and confers no rights.

Use of included script samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm
 

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