[AGENT++] EngineIdTable always replaces entry, even when host is different

[Jochen Katz]

Hi,

> 20100526.15:18:45: 5704: (2)INFO   : v3MP::EngineIdTable: replace
> entry (old id) (old host) (old port) (id) (host) (port):
> (80 00 1F 88 80 7C 4B 2B 3A 9B 72 CE 48), (192.168.21.1), (161),
> (80 00 1F 88 80 7C 4B 2B 3A 9B 72 CE 48), (192.168.27.1), (161)

it looks like the two agents use the same engine id, which is not supported.

Regards,
  Jochen

Am 27.05.2010 21:25, schrieb KSly:
> I have a scenario where the first poll of several IPs are all
> successful, but on subsequent polls, the IP on which I am running the
> manager always fails by SNMP time out, while all other IPs are fine.
> One thing I notice is that for every snmp request, the EngineIdTable
> reports that it does not know the engine id for host/port, however it
> then replaces an entry for a different IP address.  This is the log
> when requesting information from two different IPs (192.168.21.1 and
> 192.168.27.1).
> 
> Not shown here are many get requests to each IPs, all of which are
> successful.  First are requests to 192.168.27.1, then to 192.168.21.1.
> The first line is the final request to 192.168.21.1.
> 
> I see that the entry for 192.168.27.1 is replacing the entry for
> 192.168.21.1, which makes me suspicious.  Except after the polling to
> 192.168.27.1, the same requests go to 192.168.21.1, and the entry is
> replaced again.
> To be clear, the manager is running on this IP that fails (192.168.27.1).
> 
> Does anybody have any suggestion to why my subsequent requests are
> failing for the IP on which I'm running the manager?