We have a pair of ACE modules running c6ace-t1k9-mz.A2_1_1a.bin. They are a redundant pair, one in each of two 6513s.
A few weeks ago we had to reboot the active ACE as none of the load balanced services were working through it. We could access the CLI via the supervisor and it looked ok but was reporting probe failures on the rservers and we couldn't ping anything in the arp table, although the arp entries were correct, even after clearing the arp.
We switched to the standby ACE which restored service, reloaded the primary and then switched back. Everything was ok for a few weeks but we have a similar problem now. We've switched over to the standby ACE and everything is working but we haven't reloaded the other card yet.
We're seeing probe failures to rservers in each of the two contexts running on the ACE but the same rserver probes are working fine on the active ACE. We can ping a few of the rservers from the affected ACE but not all of them. The arp and mac address entries look ok.
Before we start looking at sniffer captures I'd like to know if there's anything we can check from the CLI that may give us an indication of what the problem is.
Can anyone help?