"Holes" in graphs; Graphs are choppy

[Sgurd]

I'm new to cacti and not sure how to describe this annoyance, so here's a pic:



All of my graphs look like this, but the "holes" or misses are not always at the same times.

Is cacti failing to poll the devices at these times?

I'm running 8.6.j on Ubuntu 7.10 server. 2GHz processor, 1Gig of RAM.
Was running an earlier version of cacti before and not having this problem.

Thoughts? - JD

[TheWitness]

What poller type "cmd.php" or "cactid"?

TheWitness

[Sgurd]

What poller type "cmd.php" or "cactid"?

Crontab says 'poller.php'. I haven't install cactid.

[TheWitness]

Please review/post your log around these times. Could be that you need to increase snmp timeout value.

TheWitness

[Sgurd]

Please review/post your log around these times. Could be that you need to increase snmp timeout value.

There's an awful lot of this in the logs:

Code: Select all

11/12/2007 12:00:17 PM - CMDPHP: Poller[0] ERROR: SQL Exec Failed "insert into poller_output (local_data_id, rrd_name, time, output) values (1261, 'traffic_out', '2007-11-12 12:00:17', '2270392489')"
11/12/2007 12:00:17 PM - CMDPHP: Poller[0] ERROR: SQL Exec Failed "insert into poller_output (local_data_id, rrd_name, time, output) values (1262, 'traffic_in', '2007-11-12 12:00:17', '1150413214')"
11/12/2007 12:00:17 PM - CMDPHP: Poller[0] ERROR: SQL Exec Failed "insert into poller_output (local_data_id, rrd_name, time, output) values (1262, 'traffic_out', '2007-11-12 12:00:17', '3327526894')"
11/12/2007 12:00:17 PM - CMDPHP: Poller[0] ERROR: SQL Exec Failed "insert into poller_output (local_data_id, rrd_name, time, output) values (1263, 'traffic_in', '2007-11-12 12:00:17', '1439547528')"
11/12/2007 12:00:17 PM - CMDPHP: Poller[0] ERROR: SQL Exec Failed "insert into poller_output (local_data_id, rrd_name, time, output) values (1263, 'traffic_out', '2007-11-12 12:00:17', '1345312266')"

. . . every time the poller runs. Can't say I know what that means exactly.

Thanks for you replies so far.

[TheWitness]

Sounds like a double cron issue. Verify only one is set.

TheWitness

[gandalf]

Find the issue Larry's referring to at the second link of my signature
Reinhard

[Sgurd]

Sounds like a double cron issue. Verify only one is set.

I checked the crontabs for myself, root, and cacti. Cacti was the only user w/ anything set.

Is there something else I can check? Or more info. that would help?

[TheWitness]

Maybe mysql connections? Or, maybe mysql is rebooting. You should look for some error messages there. I found a problem the other day with Key Cache above 4GBytes.

TheWitness

[Sgurd]

Well, I started using cactid yesterday around 4pm and haven't had any problems since.

Thanks for all of the replies.