What if my mail server goes down.

[Thekernel]

Is there a way to setup and extra mail forward in case my mail server goes down. We still need to be able to receive notifications if that happens. WE are using Nagios for Linux. Any ideas on what config files need to be modified or what the process would be to set this up.

Thanks

[Thekernel]

Sorry, I think I posted this in the wrong section. We are using Nagios 3.2.1. I'm very new to supporting this nagios and they just kind of threw it on me this morning, so any help would be much appreciated on how to get this setup.

Thanks!

[agriffin]

There is really no simple way to do this. The simplest way to do it is to set up each contact with multiple email addresses (using the addressx directives and the $CONTACTADDRESSx$ macros) and send notifications to both addresses all the time (not only when your primary email server goes down). That's not very convenient but your other options are much harder to implement. You could set up an event handler to change the notification command when your mail server goes down. You could also try to ensure that your mail server doesn't go down by making it part of a HA cluster.

[mrtexasfreedom]

Agriffin has some good suggestions. Another thing you could do to reduce the promiscuity of email messages is to set up an escalations email using a third-party provider. This could even be a text message to a cellphone <yourphonenumber>@sprint.net after an hour of a critical error not being resolved. Escalating through a gmail address would also work if you're watching that gmail box constantly.

The bigger threat would be if your monitoring server loses internet connectivity. In that scenario, all bets are lost unless you have an outside server monitoring the main monitoring server.

-- mtf

[tigerhead]

I was also curious about this possibility. good thing someone else already thought of it. score!

[crfriend]

This is actaully a wonderful example of why one ought to use "out of band" notifications instead of naving one's notification channel depend on something that is under monitoring and prone to failure. An example of this might be a directly-connected modem that dials out to a TAP gateway for paging, or a directly-connected SMS device that can contact mobile phones and send text messages to them.

True, this does lead to some extra complexity, but it greatly increases the overall reliability of getting notifications out of the local infrastructure and, thus, away from one's potential local problems. Even if the entire local network drops (e.g. a failure on a switch or spanning-tree insanity) or the wire coming into the building is cut the word will get out.

[GaWd]

My simple solution-which might not be practical for everyone-is to use a second instance of Nagios that uses a different SMTP server to monitor mission critical hosts and services.

If one mail service is hosed, the other is very unlikely to be. Not everyone has more than one SMTP server, but I just happen to-my exchange server and a secondary SMTP server in my co-lo. I have a main instance of Nagios in my co-lo and then my local offices have a Nagios instance as well.

[mikeyred]

thanks for the info I was having some trouble too