NLS CLuster and Alerting

[Mitchell]

Okay , i got it , yes please put that as feature request . In fail-over situation all the alerts would fire from active node is this correct ?
on a side note right now we are getting alert email like root@node1 and root@node2 , if we want to see the send email from Nagios log server Alerts <[email protected]> instead of root@node1 and root@node2 what changes we need to make ?

Thanks
Ramesh

[jolson]

on a side note right now we are getting alert email like root@node1 and root@node2 , if we want to see the send email from Nagios log server Alerts <[email protected]> instead of root@node1 and root@node2 what changes we need to make ?

You should be able to modify the 'from' address directly in the mail settings:

2015-07-30 11_53_17-Mail Settings • Nagios Log Server - Firefox Developer Edition.png

Okay , i got it , yes please put that as feature request .

I have made a feature request to specify which instance that emails come from. The task ID is 6162 - feel free to reference this ID to any Nagios team member if you'd like to check on the progress of the feature request.

[Mitchell]

Thanks , i did tried to change the from email directly from Mail settings but still it has root@node1 and root@node2 , do i need to specify "[email protected]" somewhere else in the nagios log server configuration ?

[jolson]

Were you certain to press the 'Save' button after you changed the 'From' address? I have NLS R2.1 installed and this is working well on my system.

[Mitchell]

I was not saving it , saved the changes and tested and it works fine now , is it a bug that we need to save the changes before testing new configuration or its design is like that?
Thanks a lot for your help

Thanks,
Ramesh

[jolson]

is it a bug that we need to save the changes before testing new configuration or its design is like that?

It's intended that saving is necessary before configuration changes take place. If you have any suggestions about how we could make the save process better, I would be happy to generate a feature request with your comments.

[Mitchell]

after saving changes and testing it was updated the from email address , now what i am seeing is since yesterday after 4 PM we got around 70 alerts with correct from email address but what we are noticing is all alerts has node1 link in the alert body , this means no alerts were fired from node2 if alerts were fired from 02 then in the body it would say node2 link , this is highly unlikely that all alerts were fired from one node , i am suspecting after the update yesterday somehow on node 2 alert is not working , what could be the issue ?

[jolson]

what could be the issue ?

The URL for the alert function is automatically pulled out of 'global settings' - which you have set to 'node 1'. Your mail is likely being sent from both nodes, but the URL will always be whatever you have defined in 'global settings'. This is because many people use round-robin DNS to balance the load of their Nagios Log Server cluster, and it's convenient to have a single domain name that may resolve to any IP in the NLS cluster.

[Mitchell]

Initially this is what i thought and validated by changing the node2 in global setting fired the alert and alert body has node2 link in it , the interesting this is before updating the from email address when it was default root@localhost and phpmail it was alerting from root@node1 and root@node2 and each alerts had their respective node name in alert email body like if the alert was coming from node1 alert link had node1 link and if the alert was coming from node 2 then the alert body had node2 link , other than changing from email address i changed it to SMTP server as well instead of PHP

Thanks,
Ramesh

[jolson]

I am under the impression that you didn't have the 'from address' field set until recently - if that's left unset (and you're using PHP Mail) the local server will use "user@hostname" as the from address, which would explain why you were seeing the discrepancy. I'm glad we got this all figured out!