Trying to use cactid instead of cmd.php

garybob · Post by **garybob** » Thu Aug 04, 2005 5:59 pm

Hi,

I am using cacti 0.8.6c with php 5 .0.3and net-snmp 5.1.2. It works
great with cmd.php but runs about 164 seconds with 207 hosts.

cactid 08.6e was installed and I attempted to use it. It runs and works great on about 95 percent of my hosts but on the others it will not update
the rrd files. It flags these hosts as "down". All the ones in which I am
trying to collect wireless associations go into this state as well as as about 10 Cisco routers from which I am attempting to collect interface data, ping times and memory usage. The other 95 percent that are working are also Cisco routers - same data is being collected.

I tried clearing the poller cache, etc.

When I switch back to "cmd.php" the "down" devices go into a "recovering" state and then after about two polls go into an "up" state,
and data gets collected again!?

Can anyone shed a light on what is happening here?

I turned on debugging with cmd.php and also cactid.
When cmd.php runs, I can find the hosts that are not being collected
with cactid. But with cactid these hosts do not show up in the log;ie
devices cap600-1a, cap600-1b thru 1f.

Post by **TheWitness** » Thu Aug 04, 2005 6:43 pm

You have change your availability method to SNMP Ping only. There is a known issue in 0.8.6e. I just don't have the fix for (In addition to any *ix boxes) it and no one discovered it during testing.

TheWitness

garybob · Post by **garybob** » Fri Aug 05, 2005 3:26 pm

Thank you very much! This fixed the problem.

The statistics are pretty amazing.

08/05/2005 04:07:50 AM - SYSTEM STATS: Time: 170.0071 s, Method: cmd.php, Processes: 15, Threads: N/A, Hosts: 207, Hosts/Process: 14

08/05/2005 10:45:24 AM - SYSTEM STATS: Time: 22.7678 s, Method: cactid, Processes: 5, Threads: 10, Hosts: 207, Hosts/Process: 42

I believe I saw a Metrics F0rum. I will post this information there also.

Cacti