Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
My environment is QV 11 SR2 with 3 dedicated physical servers. QVDS is on server 1 while QVS is on servers 2 and 3. QVS is clustered on server 2 and 3.
We have noticed that every 3-4 weeks (on average) QVDS and QVS hang and must be restarted. My server guy implemented a script that automates the restarting of QVDS and QVS on a weekly basis. After the first run of this script I noticed that my QVS’s took roughly 12 hours to come back up (each). This caused obvious issues since QVDS tasks failed attempting to distribute to my QVS’s which were both down at the time. When I restart QVS manually the services come back online in just a matter of minutes.
So, for those of you who have implemented automatic restarts of QVDS / QVS – are there any tricks or best practices? One question – does it matter what order your restart your services; specifically the QVS cluster master? Should I be doing a physical restart of the server itself, or just the QVDS/QVS services? I do run IIS, so when I restart QVS should I be restarting IIS as well?
Thanks in advance for any advice.
- Trevor
Hi Trevor,
Picking up on this thread, what did you guys used to create the automation scripts ?
your respond would be much appreciated, thanks.
This can be done with a Windows Tasks and a Batch file using:
net stop QlikViewServer
net start QlikViewServer
This will stop and restart the QlikView Server depending on the Windows Task.
It works.
Thanks !
Hi
Taking 12 hrs to come back up is not normal. Check your windows logs to see what is going wrong. We restart every night with no such issue. I would definitely recommend to upgrade to SR9. Check the bug list before you do so.
Hi,
we are just planning to implement that, too, after we had mayor QVS problems last July or so - we just updated to QVS 11 SR9 because of this, among other reasons.
It was a mayor thing here to convince the IT_guys of the sense of having such a thing ...
Our external consultant from KS Solutions, a partner of CP, told us that only one of the services actually needs to be restarted - the DistributionService. No need to worry about the others. We haven't tested this, though.
Best regards,
DataNibbler