High cpu utilization on vrrp-cluster

[mmuessig]

Hello forum,
i am in trouble with an active/standby setup (two nokia ip 130 with vrrp) and NG R55 HFA_18.
After a switch reboot (all fw-ports are in portfast-mode) load on both members suddenly raised to 99%.
On a ssh-shell the following two processes seem to cause the load-problem (same on both clustermembers):

root 9443 42.3 8.5 15464 21460 ?? Ss 9:31AM 15:01.23 /opt/CPfw1-R55/bin/cphamcset
root 21142 32.7 2.5 612 6432 ?? S 10:24AM 0:02.04 /opt/CPfw1-R55/bin/cphaconf macs

I tried to restart cpha (cphastop followed by cphastart) but without success.
Also a cphamcset-process survives the cphastop, why?

Any ideas where the utilization comes from?

Kind regards

[crisbert]

Hi,

I'm experiencing the same behaviour, have you found a solution?

We have here 4 clusters configured in the same way (and at the same software release). They were all suffering of the same problem, an update to HFA_17 solved the issue only on three of them.

Installing HFA_17 on the last cluster, it made vrrp stop working (uninstalling the hotfix vrrp starts working again).

any idea?

Thanks

Cristian Bertoldi

[northlandboy]

Try running cphaprob list, to find out why VRRP is stopping.

I have seen something similar to this when multiple clusters under one management server had the same sync network configured on them.

Also check the names of all your interfaces in Check Point exactly match what's on the modules. Turn off Automatic Proxy Arp configuration, if you've got that enabled.