AIQ: Invididual used%-alerts per volume

[Tava]

Hello,

I know global setting for used% is 80% nearly full and full at 90%. Now I would like to set higher for one test-volume, so here goes (AIQ 9.7P1):
 -Test-volume "used%" is currently 96%
-Storage->Volumes, mark my testvolume -> Edit Threshold -> Capacity: Object Level -> mark -> set:
  -Nearly full: 90%
  -Full: 95%
-Save&Close
->Storage Management -> Alert setup -> Add 
  -Name: "Testing vol-used"
  -Resources: <selecting my test volume> move to right column
  -Events: Error -> volume space full, volume logical space full, volume space nearly full and volume logical space nearly full
-Actions -> set my email and notification "remind every 15 mins"
After saving, I check that my alarm is "enabled".

I press "Test" and get email saying "test email etc" from AIQ. This is all I ever hear from this alert. My vol is over "error-level" 95% so how come I don't ever get a email about this???

Also I don't get events about this on "Event Management". No individual nor global email.

I should have both, right?

-Jari

[helge]

If it is not full but nearly full it a warning, not an error.

The error would only show after polling. In a test scenario you need to fill the volume, discover the cluster and then the alert should come.

If you mark it as resolved it should be resolved. The alert will not be triggered again if the error condition remains.

For each try you need to empty the volume, discover- so you get the obsolete event and then fill it again and discover.

[Tava]

Thanks for your answer! So based on that, you mean I cannot set alert for volume which has already breached my alert setting?

If my volume is already 96% and the alert setting is 95%, it won't trigger the alert? In order to get the warning, I need to get used% under 95 and when it breaches the limit, it will trigger the warning?

[helge]

Yes to the first question. Alerts are based on events. If the event was triggered and you set an alert it will not be send in already breached condition.

Second one depends. If you add a cluster in that condition, I twill alert you immediately.

If you had an alert which you cleared, It will not alert again.

Yes, if you fix the problem, do acquisition (re-discover or wait 20min) that UM is aware it is fixed- if it happens again it will trigger the alert with next polling/data acquisition.