I recently opened a support case for best practice for monitoring networker (what processes, what alerts to generate, why the heck a tape marked as cleaned is a critical alert that gets me paged at 2am), and got nothing useful from the support team, so I through it out to everyone.
What do you monitor? What alerts/notifications do you have Networker generate?
We are trying to generate an alert for when the Networker server can't reach the Storage Node (either network issue or nsrlcpd hung) and yet Networker doesn't appear to generate a notification for this event, despite happily sending me 300+ critical alerts every day for tape changes/cleaning.
We are running 7.4.4 on Solaris 10. (Yes I know, we need to upgrade).
Well, if you have 300+ critical messages per day perhaps you may wish to focus on those first.
I monitor only for failed groups - that's all that matters.
We have daily check mechasnim for available volumes in pool so we never get an issue with tapes being waited for. On top of that 95% of tape drives we use are virtual ones so no cleaning messages either. We use i500 as PTL and I never had cleaning request yet in NW (but then again, we do not allow software application to handle cleaning of HW which HW can handle itself).
Availability of sn was never an issue, but I would never monitor that either as that would be picked up by monitoring tools for basic network issues or host itself (we use HP OpenView).
I figured as much. The 300+ alerts aren't actually critical, they are MEDIA NOTICE alerts, which is the same level alert if a SN goes offline. The exact problem we had was that the monitoring system was able to connect to the Storage Node on the correct ports, but the NW server couldn't. So the only place we can monitor this from is the Server.
Why a SN going offline is only deemed a NOTICE level alert is beyond me.
Why do you use HP OpenView for monitoring? What about DPA?
Are you not an EMC employee?
I use HPOV as this is our monitoring tool for whole DC. DPA was not something I found working nice in our env (it did collide with certain setups we had so despite nice features it had to go).
I'm not EMC employee. EMC employees on these forums are usually easily recognized as they have icon next their name indicating they are EMC staff. The icon next to my name indicates subject matter expertize in certain area of EMC world, but not being EMC employee.
What type of alerts can you send to DPA from Networker? Everything that you see in group status, ie error/warning/critical etc? Can that be sent to BMC/HPSM to generate tickets? Any KBs or whitepaper on this..
Yes, you can use DPA to send traps to your ticketing tools.
You can configure DPA to gather certain reports and then these can be sent to your tool.
Some information can be found in the following links, however I think the documentation will give you much more information:
Although these links will point you to contact EMC support.