UNSOLVED
MG
Major Goat
1 Rookie
•
14 Posts
0
1612
January 9th, 2008 08:00
Test firing alerts
Hi all,
This is quite easy with quantitive alerts such as filesystems x% full - just lower the threshold, set to run every couple of minutes and wait.
Is there a similar way to do this for qualitative alerts ( True/False ) conditions ?
We are integrating CC with Tivoli and we want to test Tivoli's response when it receives an SNMP trap following an event such as a failure of a disk, FA etc.
Ideally, there would be a nice "Test Fire" button somewhere, but without that, what else would you suggest?
Happy New Year to everyone,
Nick Buckley,
Cardiff,
UK
This is quite easy with quantitive alerts such as filesystems x% full - just lower the threshold, set to run every couple of minutes and wait.
Is there a similar way to do this for qualitative alerts ( True/False ) conditions ?
We are integrating CC with Tivoli and we want to test Tivoli's response when it receives an SNMP trap following an event such as a failure of a disk, FA etc.
Ideally, there would be a nice "Test Fire" button somewhere, but without that, what else would you suggest?
Happy New Year to everyone,
Nick Buckley,
Cardiff,
UK
744
0
Posted January 29th, 2008 13:00
Hi Nick,
If RRR's post answered or helped answer your question, please mark "Answered" or "Helpful" to award RRR the points.
Thanks,
David D. Ward
EMC Resource Management Software Group
Forums Moderator/Host744
0
Posted January 30th, 2008 00:00
You can also play around with the numer of times a threshold should be exceeded before a "trap" is generated (or cleared again). Works great for me !
This particular excercise is covered by the ECC course. Great course... 5 Days of EVERYTHING about ECC......744
0
Posted February 26th, 2008 02:00
Agreed,
This is a great way to reduce alerts due, for example, to transient spikes in performance.
However, I this isn't what I meant.
Thanks nonetheless.744
0
Posted February 26th, 2008 02:00
Hi,
That's the correct strategy for 'continuous' alerts e.g. filesystem x% full, but I don't see that this would ever work for a 'discrete' alert e.g. FA failed. That's either true of false and not at all easy to emulate or trick ECC into believing it should fire the relevant alert.744
0
Posted February 27th, 2008 07:00
I'd love to hear the answer on this as well. We are trying to integrate with HP OV to do some event correlation and paging out of PowerPath "All Paths Dead" alerts. So far the only way we have been able to "simulate" is to set up a test host and yank all the storage out from under it... then give it back... then take it away again...
Very time consuming for testing.744
0
Posted September 22nd, 2011 02:00
Hi everyone,
Well, 2 and a half years later, we finally have a working solution.
We installed the Storage Agent for Symmetrix on an AIX host and assigned six Gatekeepers too.
Note that the Master Agent and Host Agent for AIX were already installed.
Thereafter we set two or three GKs to Not Ready and waited a few minutes ( the Symm Agent has a 2 minute interval set for the Alert Polling DCP )
The Alert fires as expected and is displayed on the console.
The attached Management Policy also works and duly sends a confirmation e-mail.
Regards,
Nick
Note: don't forget to ready the GKs afterwards
RRR
6 Operator
•
5739 Posts
744
0
Posted January 10th, 2008 07:00
(But do make sure you don't create an avalanche of traps, because the threshold is too low or right on the edge of being crossed every minute or so)