I’m going to be lab testing OSPF today on a few Soekris boards. In the near future I hop to implement this across our network for backhaul redundancy of sorts.
One question first that comes to mind and I can’t seem to find a decent answer is this:
If a link were to fail and be re-routed via OSPF, does OSPF have means or capability of notification other than LSA updates etc… Meaning; Hypothetically speaking if TowerB was a backhaul to TowerA and TowerB’s backhaul went down to TowerA and OSPF rerouted TowerB to TowerC (which TowerC has a link to TowerA); What means of notification could there be? Other than routing table changes. It’s like RAID…it’s there for redundancy, but at some point of fault/failure, there needs to be a notification of this. I can’t see a simple ping as being helpful because it will just be re-routed and make it through.
What methods do you use for redundancy?
Couldn’t you just use Nagios to monitor the Private IP of the BH’s?
If a BH loses power then you’ll get an alert from Nagios because if won’t be able to ping it anymore.
i am not faimilar with Soekris boards, but are you looking for failover in which case have you considered spanning tree, if it is both failover and loadbalancing then I guess OSPF.
To answer your question OSFP will update the cisco log if you have it switched on, which can be configured to send the log to a syslog server, which you can poll using some sendtrap software…
however the easiesy way of monitoring is to use Frothingdog suggested a monitoring tool, Nagios is a good one, if you are familiar with Linux, others are Intermapper (pay) opManger, and many more… also depends what you want to do if something goes down, is someone going to be monitoring all the time and you need a visual display or do you want an SMS/email sent out, again these applications can configured accordingly
I was looking for sms/paging capabilities and believe I have a method for invocation now.
Thanks for the replies folks.
we use RSTP with PVST and has been rock solid so far.
For SMS, you can tryout Nagios+kennel , this is what we are using since last 6 years.