Conference noted::netview

2225.0. "What timeouts are used when status polling is done due to receiving a trap" by SWETSC::TORLIND_L () Mon Mar 24 1997 16:29

	
	Hi,
I have a question regarding status polling timeouts. According to the man pages the netmon polls a device and all
it's interfaces immediately after receiving a trap from that node. Is the configured timeout values used or are
there other default's used. 

I have set up a global polling record of 10 min between polls and a timeout of 60 seconds and 3 retries.

I have seen several times when there have been shorter disrupts (30-60 sec) that some nodes turn red. Why ??


One thing I tried to day was the following, between two sites there are two wan connections from two different
routers. I killed the primary path and was doing ping's and after 15 sec the secondary path was used.

And again there were a couple of nodes that went red. Why ?? I thought that when I have 10 min between polling,s
and a timeout of 60 sec's and 3 retries I would survive these short interrupts. All help will be appreciated.

/Lars-Ake Torlind Stockholm Sweden

If you want to see what timeout values is used by status polling.
you should set trace mask to netmon.

# netmon -M 3
           ^^^3 means 1(ICMP echo request) + 2(ICMP echo reply)

Next you will do same test that you did,
and stop trace.

# netmon -M 0
           ^^^turn off trace

Last, you should see trace log file "/usr/OV/log/netmon.trace"

14:54:30 :sending ping to 16.162.48.84
               seqnum = 34570 ident = 31071 timeout = 11 
                                                      ^^ here
14:54:41 : expired ping to 16.162.48.84 (zotto.mix.ewb.dec.com)
               seqnum = 34570 ident = 31071

If you want also to see a relation about reciving trap and status polling,
add trace mask +32(Trap receive).


--
Yoshiharu Enya

Thank's for your response, I'll try that.

/Lars-Ake Torlind