Subagents lose connection to master if no reply from inform destination

I am running into an issue where several subagents lose their AgentX
connection to snmpd master, all giving the following error message:

"AgentX master agent failed to respond to ping.  Attempting to re-register"

This occurs after they successfully connect to snmpd.

I have two v3 inform destinations configured using the trapsess
derective in snmpd.conf:

trapsess -v 3 -r 2 -t 15 -u <user> -l authNoPriv -A <pass> -a MD5 -Ci
10.11.10.18
trapsess -v 3 -r 2 -t 15 -u <user> -l authNoPriv -A <pass> -a MD5 -Ci
10.11.10.19

The second one I eventually discovered, was un-reachable.

When I removed this inform destination. the subagents did not lose
their connection. Also, if I set the retry & timeouts to 0 for the
un-reachable inform destination, the subagents do not lose their
connection.

So, I am left with the conclusion that some how having a "bad" inform
destination effects the agentx connection between master and
subagents.

I encounter the same issue whether I configure to use Unix domain
sockets, or TCP sockets for the AgentX connection.

I realize that having an "un-reachable" trap destination configured is
not ideal. But, I would like to avoid having this effect the subagents
if possible.

Any help is greatly appreciated.

System:
Debian Linux, Net-Snmp 5.4.2

John

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables unlimited
royalty-free distribution of the report engine for externally facing 
server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Net-snmp-users mailing list
[email protected]
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users