network usage problem

[Jeff Hagstrom]

I have an RP7410 running 11.i and we've noticed a slow down on our network in the last 4 weeks. Some software that our network guy has, points to the RP box. Every 20 minutes there seems to be something that spikes for a few minutes then goes away for 20 minutes. Do you have any idea what to look for, where to look for, any free products that might narrow down where the problem is, any information at all.

[TTr]

If I understand you correctly, these spikes on the rp server cause the slow downs on your entire network?

Start by looking at cron jobs. Then go on to look at a full "ps -ef" listing and examine each process. At the same time think of any agents that are running on your server that could be connecting to a central collector and sending information out. These agents could be performance monitoring and statistics and could be for the OS, the database or the app.

[Rasheed Tamton]

The first thing I will look is in the /var/adm/syslog/syslog.log file. The 20 minutes interval looks like, perhaps, is caused by the syslogd.

[Laurent Menase]

20 minutes it the ip_ire_flush_interval
which means that every 20 minutes routes to destinations are reevaluated.


You probably have more than 1 default gateway
and one points to a dead gateway.
or arp or ping requests to the default gw doesn't get answers soon enough.

You may check ndd -get /dev/ip ip_ire_status for dead gateways.
you may need to take a nettl trace and stop it as soon as you see the problem.
You may also contact hp support. - they will ask for at least the 2 info above.