Cisco Prime Alarms not cleared

[steiostb1]

Prime is generating too many alarms. Critical alarms which is not active is still listed as active. For instans, when an AP reboots -  an critical alarm is generated. But when the AP is up again, the critical alarm is still listed under critical alarms in the dashboard startpage. This is from the event log for the AP: 

AP 'AP7c0e.ce48.4ff8' disassociated from Controller '. Critical

Access Point 'Ap' associated to controller '. Reason for association 'None, Information

This shows that Prime is received the auto-cleared alarm from the device, why is still the ciritical alarm showed in the dash? In WCS the alarm is cleared and not showed in the dashboard.

.

[steiostb1]

Is this a bug or is it a setting for alarms to be cleared? I have checked in Administration - settings - system settings - alarms and events.

According to Prime 3.0 user guide, Alarms should be auto cleared:

Recurring Alarms and Events

To reduce the amount of unnecessary alarms and events, Prime Infrastructure detects the underlying causes of an event, and then modifies when it issues alarms and events if the devices have any of the problems. For example, for module or link faults, if a module is down, Prime Infrastructure creates one Module Down alarm only, and associates all of the interfaces’ link down events to the Module Down alarm. When the module state is restored, Prime Infrastructure clears the module alarm and all interface messages are associated to the cleared alarm.

[nalway]

I'd have a chat with TAC if I were you. At least this will raise the profile.

Have the same issue (amongst others).

TAC still do not know what is causing it, not even recognised as a bug.

EDIT: Database was screwed due to patch .51 introducing elements that did not interoperate with originally installed database. Cisco recognise this is a fault and may be able to resolve if no other remediation works have taken place.

[steiostb1]

nalway: Any updates from your TAC-case? What controller are you using?

[nalway]

Server has just been rebuilt from scratch yesterday.

TAC said that patch .51 (and also .53) added things to the database which effectively broke it. Only way to resolve with confidence was to rebuild. Not happy but at least the server is working again.

The issue regarding alarms was due to patch .51. I'd recommend getting in touch with TAC and have them review.

[steiostb1]

Thank you for your reply!

What do you mean by patch .51?

sh version outputs this:

Version : 3.0.0
Build : 3.0.0.0.78
Critical Fixes:
PI 3.0.2 ( 2.0.0 )
PI 3.0.2 Update 02 ( 2.0.0 )
Prime Add-Ons:
Prime Insight Agent ( 1.0.0)

[nalway]

Take a look here regarding the mother of all cock ups so far in the life of PI 3.0 :

https://supportforums.cisco.com/discussion/12733941/prime-infrastructure-30-problem

Looks like your patching is ok, possibly short of the device pack which was a recommended install for our server. Check with the wonderful (...) TAC team if you are in doubt :

Version : 3.0.0
Build : 3.0.0.0.78
Critical Fixes:
        PI 3.0.2 ( 2.0.0 )
        PI 3.0.2 Update 02 ( 2.0.0 )
Device Support:
        Prime Infrastructure 3.0 Device Pack 3 ( 3.0 )
Prime Add-Ons:
        Prime Insight Agent ( 1.0.0 )

[steiostb1]

Thanks again.

The problem seems to be with the update_02-1.0.7.ubf, - https://supportforums.cisco.com/discussion/12733941/prime-infrastructure-30-problem#comment-11122861

Cisco have already reported it as a bug, but its internal and not open for us to see: CSCux31956 Alarm severity change not getting updated for alarms in DB.

[nalway]

Hey - take a look back at that thread : https://supportforums.cisco.com/discussion/12733941/prime-infrastructure-30-problem

Seems like there is another bug regarding alarms : CSCuy28245

[nalway]

All

Cisco have (or will be very shortly!) released a patch for this.

I don't see it on the Patch area of the PI 3.0 downloads and it hasn't been detected by my PI server but I will have a WebEx with them to install it today.

[Ditter]

Hi to all,

Found this old post, but it seems that this problem is not resolved even in Ciscp Prime 3.9

Critical Alarms that are not getting cleared automatically when the condition is not valid. 

For example , during a power outage i got a critical alarm for "redundant power supply failure".

This critical alarm was in category switches and routers and the condition was " Environmental Monitor Redundant Issue"

However this alarm appeared 4 days ago and the alarm was cleared after 1 hour in the same day. But after 4 days i am still having this alarm active. Shouldn't it be auto cleared by itself?

Thanks,

Ditter.