[AGENT++] Duplicate Engine IDs?

[Jochen Katz]

Hello,

> obviously!). Duplicate Engine IDs seem like a very bad thing to me
> but this seems to be how Cisco set them up ...

yes, really bad thing...

> When we try to talk to these devices we can successfully query one of
> them but the queries to the other always fail with a timeout.

Most likely the engienBoot and engineTime values are not the same for
the two agents. This results in dropped requests/responses and you see a
timeout.

> The process doing this is long running so I am thinking that maybe
> something is caching the engine ID? A test using standalone tools
> like NET-SNMP snmpwalk always works with both devices

Sure, these do not cache the time information.

> Anyone have any clues!?

You can use the following workaround:
Call the following function to remove the time counters from USM:
int USM::remove_time_information(const OctetStr &engine_id)

Regards,
  Jochen