Centos 5 errors

[digitlman]

New install on Centos 5 x86. Everything was fine until yesterday, after I added a few more devices. Now I get this:


09/09/2010 07:55:14 AM - SYSTEM STATS: Time:13.4022 Method:cmd.php Processes:1 Threads:N/A Hosts:15 HostsPerProcess:15 DataSources:707 RRDsProcessed:357
09/09/2010 07:55:01 AM - POLLER: Poller[0] WARNING: Poller Output Table not Empty. Issues Found: 2, Data Sources: traffic_in(DS[45]), traffic_out(DS[339])
09/09/2010 07:50:16 AM - SYSTEM STATS: Time:14.4273 Method:cmd.php Processes:1 Threads:N/A Hosts:15 HostsPerProcess:15 DataSources:707 RRDsProcessed:15
09/09/2010 07:50:16 AM - SYSTEM STATS: Time:14.4547 Method:cmd.php Processes:1 Threads:N/A Hosts:15 HostsPerProcess:15 DataSources:707 RRDsProcessed:370
09/09/2010 07:50:15 AM - CMDPHP: Poller[0] ERROR: A DB Exec Failed!, Error:'1062', SQL:"insert into poller_output (local_data_id, rrd_name, time, output) values (403, 'traffic_in', '2010-09-09 07:50:14', '54319144')'
09/09/2010 07:50:15 AM - CMDPHP: Poller[0] ERROR: A DB Exec Failed!, Error:'1062', SQL:"insert into poller_output (local_data_id, rrd_name, time, output) values (403, 'traffic_out', '2010-09-09 07:50:14', '153470309')'
09/09/2010 07:50:15 AM - CMDPHP: Poller[0] ERROR: A DB Exec Failed!, Error:'1062', SQL:"insert into poller_output (local_data_id, rrd_name, time, output) values (404, 'traffic_in', '2010-09-09 07:50:14', '706497')'
09/09/2010 07:50:15 AM - CMDPHP: Poller[0] ERROR: A DB Exec Failed!, Error:'1062', SQL:"insert into poller_output (local_data_id, rrd_name, time, output) values (404, 'traffic_out', '2010-09-09 07:50:14', '67444173')'
09/09/2010 07:50:15 AM - CMDPHP: Poller[0] ERROR: A DB Exec Failed!, Error:'1062', SQL:"insert into poller_output (local_data_id, rrd_name, time, output) values (405, 'traffic_in', '2010-09-09 07:50:14', '135241120')'
09/09/2010 07:50:15 AM - CMDPHP: Poller[0] ERROR: A DB Exec Failed!, Error:'1062', SQL:"insert into poller_output (local_data_id, rrd_name, time, output) values (405, 'traffic_out', '2010-09-09 07:50:14', '217965056')'
09/09/2010 07:50:15 AM - CMDPHP: Poller[0] ERROR: A DB Exec Failed!, Error:'1062', SQL:"insert into poller_output (local_data_id, rrd_name, time, output) values (406, 'traffic_in', '2010-09-09 07:50:14', '0')'

It looks like one source in particular. Should I just re-create this device?

[TheWitness]

Looks like the double cron issue. Make sure only one poller is running.

[digitlman]

I ran "crontab -e" and there is only one line for cacti.

[TheWitness]

Did you install from a package? If so, they typically place entries in /etc/cron.daily or some other location under /etc.

TheWitness

[digitlman]

Yeah, from a package.

/etc/cron.d has a file called cacti that calls the 5 min poller.

crontab -e has the stated 5 min poller as well.

Is crontab ref'n the entry in cron.d, or is it a duplicate?

[TheWitness]

You should remove the one from "crontab -e" and use the entry in /etc/cron.d instead.

TheWitness

[digitlman]

Well, that seemed to take care of the errors, but the graph still isn't updating.

Should I just re-create that device from scratch?

[TheWitness]

Run the following from the command line as the user:

Code: Select all

php -q poller.php --force

You should get a bunch of "OK" messages.

Review the debugging section in my signature.

TheWitness

[digitlman]

Ah ha!

It seems, somehow, the RRD files for that device got an ownership of root, not the cacti user.

Not sure how that happened.

Changed them to cacti:cacti and looks like we're back in business!

[digitlman]

This happened again randomly to a new graph I just created.

It this a bug with 8.7g or a cacti-specific issue?

[TheWitness]

Make sure your stats remain under your poller interval.