Several months after upgrading our AP1230G units to IOS 12.3(8)JA2,
we learned that they were intermittently failing to forward
broadcast (and multicast) frames from the wired to the wireless side.
The wireless customer typically notices it once wired devices (including
the IP router) can no longer get broadcast IP ARP requests to reach the
wireless client. As you can imagine, this causes widespread problems for
our wireless clients.
This is filed as bug ID CSCsg26708.
Previously we'd been running 12.3(7)JA1, 12.3(7)JA2, and 12.3(7)JA3,
where we didn't experience this problem. Re-testing the identical
configuration in 12.3(7)JA3 appears to show that the issue wasn't
present in 12.3(7)JA3.
I don't see any apparent pattern to when the APs get into this state,
or when they come out of this state. The problem duration can range from seconds
to minutes to hours at a time. There's no periodicity that's apparent to me.
There's no apparent synchronization among the hundreds of APs experiencing the
problem (all are attached to the same Ethernet VLAN/IP network); each AP
experiences the problem independantly of the other APs.
I've not been able to determine what triggers the AP to get into (or out
of) this state; for example, logs don't indicate it happens when a client
associates or disassociates. So although it's happening many times
per day on many APs, I don't know how to reliably trigger it.
If anyone else is experiencing this issue,
I'd love to hear if you've been able to determine what triggers it.
Knowing how to reliably reproduce it might be useful information I could
provide to cisco for the bug report.
