OSPF issues on SPLAT.

[deltaforce98]

NGX R65 SPLAT hfa_25:

We are experiencing strange OSPF issues on one of our clusters running OSPF with about 900 routes on SPLAT. After successful running OSPF for couple of hours the SPLATs loose their routes. We have tried this twice and each time the routes are there and everything via the cluster works fine but after 4-5 hrs, the routes disappear.
We have other clusters on SPLATs that also run OSPF with no issues.
Has anyone else experienced this issue?

[melipla]

I've had a problem where one interface lost adjacency & therefore lost all of its routes which originated from that side of the adjacency. It sounds like your problem is more serious--does GateD stop running altogether?

PS did you mean R55 hfa 25?

[chillyjim]

Haven't seen this myself, and I have several customers using OSPF.

Are you seeing anything in the logs?
Is it possible you don't have enough memory for the routing table? 900 is a good size routing table for a gateway. I've got one customer that in this general area (800+) but they do have gateways with 2GB of ram.

You can run gated like this:

gated -a "-tall /var/log/gated.log"

and get a lot of debugging info