Notifications cached or delayed

[Jwhalen]

Hi All,

We are running nagios XI 5.2.5 on centOS 7 and have an odd email notification issue. Any service check notification that goes out of the system has the same service description, regardless of the actual service.

As example, this service check as shown in the Notifications window is for memory usage on a windows host and was sent at 16:58:

2016-05-24 16:58:31 windowshost.xxxxxx.com win_memory_usage_phy No WARNING jwhalen

However, the email I get is for an ssl check on localhost with a date received 28 seconds later, but the error in the cert check is for 52 days ago:

WARNING - Certificate *.xxxxxx.com expires in 42 day(s) (Sat 14 May 2016 12:45:00 AM PDT).
Date/Time: 2016-05-24 16:58:59

The expires in 42 days number does go down each day as well. Yesterday it said 43.

I've checked the system time on the server, deleted the ssl check, deleted the localhost host and all checks, switched from SMTP to sendmail, restarted the nagios processes with retain status off, and ran the repair databases script. All with no change.

Also, the service check is supposed to go to 3 different addresses, but all 3 come to me. I can see them in the maillog on the nagios server all going to me, yet the notification screen shows it going to the 3 addresses, so I don't think it is a problem with local sendmail or our mail relay. I have confirmed the contact email addresses are all set correctly.

The email send test works fine and is timely with the correct information and goes to whomever is set to receive it.

I'm guessing there is maybe a notifications cache that got off track, but don't know where it is. I checked the objects.cache file and don't see the localhost or ssl check in there or referenced. I've checked the local mail queues as well and they are clear.

Is there a way to clear or remove any notifications that have been cached or queued up somewhere, or any other thoughts on a fix?

Let me know if I can provide more info.

Thanks,
John

[ssax]

I've seen a similar issue to this that was resolved by clearing out the xi_events table.

Is this a fresh install of XI 5+ or was it upgraded from a pre-XI 5 system?

[Jwhalen]

Thanks for the reply. It was a fresh install on a VM about 3 months ago, but just starting to be worked on now. The only host it had in it was localhost. We just started adding hosts, services and contacts to it last week. I did find out that another admin changed the system date about a week ago to use UTC instead of local time, but from what I can tell, that is the only change since it was built. I'll check the xi_events table to see what's in there and clear it to see if that helps.

[Jwhalen]

Clearing out the xi_events table did get rid of the old notifications that were getting sent out, but new notifications for current service checks still were not getting delivered. I cleared out the xi_meta table as well and the seems to have done it. The last few current notifications were delivered where they were supposed to go. This can be closed. Thanks!

[rkennedy]

Great to hear - closing this one out.