Cacti no reports "MySQL: 'Duplicate entry'

[CenJo]

Hi all cacti users
I am monitoring since January my servers with cacti, all seems to work well. Now I have the following problem since tow weeks.

Cacti dos stop after a couple of days updating the rrd-files, and in the cacti log there are only the messages:

04/06/2007 08:14:28 AM - CACTID: Poller[0] ERROR: Problem with MySQL: 'Duplicate entry '988-traffic_in-2007-04-06 08:14:28' for key 1
'
04/06/2007 08:14:28 AM - CACTID: Poller[0] ERROR: Problem with MySQL: 'Duplicate entry '152-cpu_user-2007-04-06 08:14:28' for key 1'
04/06/2007 08:14:31 AM - CACTID: Poller[0] ERROR: Problem with MySQL: 'Duplicate entry '16-proc-2007-04-06 08:14:28' for key 1'
04/06/2007 08:14:31 AM - CACTID: Poller[0] ERROR: Problem with MySQL: 'Duplicate entry '295-traffic_in-2007-04-06 08:14:31' for key 1
'
The table poller_out put is huge.
I can make cacti running a gain when I disable the cactid truncate the table poller_output enable cactid and rebuild poller cache. But after a couple of days the cacti is in the same state.

In the cacti log I have every 30 minutes a error message when I have 1 cactid 1 thred,

04/10/2007 08:13:56 AM - THOLD: Poller[0] Checking Thresholds
04/10/2007 08:14:59 AM - SYSTEM STATS: Time:57.1601 Method:cactid Processes:1 Threads:1 Hosts:50 HostsPerProcess:50 DataSources:604 RRDsProcessed:458
04/10/2007 08:14:59 AM - THOLD: Poller[0] Checking Thresholds
04/10/2007 08:14:29 AM - CACTID: Poller[0] ERROR: Problem with MySQL: 'Duplicate entry '988-traffic_in-2007-04-10 08:14:29' for key 1'
04/10/2007 08:14:29 AM - CACTID: Poller[0] ERROR: Problem with MySQL: 'Duplicate entry '152-cpu_user-2007-04-10 08:14:29' for key 1'
04/10/2007 08:15:26 AM - SYSTEM STATS: Time:56.8135 Method:cactid Processes:1 Threads:1 Hosts:50 HostsPerProcess:50 DataSources:604 RRDsProcessed:1078
04/10/2007 08:15:26 AM - SYSTEM STATS: Time:24.5231 Method:cactid Processes:1 Threads:1 Hosts:50 HostsPerProcess:50 DataSources:604 RRDsProcessed:158
04/10/2007 08:15:26 AM - SYSTEM STATS: Time:56.7802 Method:cactid Processes:1 Threads:1 Hosts:50 HostsPerProcess:50 DataSources:604 RRDsProcessed:993

when I make 4 cactd 5 threds I have a more errors every 30 minutes.


I have not see any duplicate poller, and I have update the php.ini memory to 64MB.

My environment on a ESX 3.0 Server:
Cacti Version: 0.8.6j
Poller Type : cactid 0.8.6i
Sever : Linux 2.6.9-34.ELsmp RedHat Enterprise AS 4.3
Web Sever: Apache 2.0.52
MySQL : 4.1.12
RRDTool: 1.2.15
SNMP Version: 5.1.2

Regard

[gandalf]

...
The table poller_out put is huge.
..
04/10/2007 08:15:26 AM - SYSTEM STATS: Time:56.8135 Method:cactid Processes:1 Threads:1 Hosts:50 HostsPerProcess:50 DataSources:604 RRDsProcessed:1078
04/10/2007 08:15:26 AM - SYSTEM STATS: Time:24.5231 Method:cactid Processes:1 Threads:1 Hosts:50 HostsPerProcess:50 DataSources:604 RRDsProcessed:158
04/10/2007 08:15:26 AM - SYSTEM STATS: Time:56.7802 Method:cactid Processes:1 Threads:1 Hosts:50 HostsPerProcess:50 DataSources:604 RRDsProcessed:993
...
I have not see any duplicate poller, and I have update the php.ini memory to 64MB.

All of this lead me to the assumption, that this is the first "triple poller" issue. Cacti will print SYSTEM STATS only at end of polling. Now it shows 3 SYSTEM STATS for one single timestamp.
Please check all of your crontabs (/etc/cron.d/cacti, /etc/crontab, cortabs of users root and cactiuser, ...) carefully. Only one single poller entry should be left.
Reinhard

[TheWitness]

TheWitness

[CenJo]

Sorry for the delay, there is no duplicate poller, but it seems to by a problem by the time synchronization.

I tried :

RH EL 4 (AMD64) / RH EL 4 Update 3 (AMD64)
AMD PowerNow! is not supported on Red Hat Enterprise Linux 4 at this time.
AMD PowerNow! must be disabled on all systems. The following steps must be followed to disable AMD PowerNow!:

[CenJo]

Sorry for the delay, there is no duplicate poller, but it seems to by a problem by the time synchronization.

I tried :

RH EL 4 (AMD64) / RH EL 4 Update 3 (AMD64)
AMD PowerNow! is not supported on Red Hat Enterprise Linux 4 at this time.
AMD PowerNow! must be disabled on all systems. The following steps must be followed to disable AMD PowerNow!:

chkconfig --level 12345 cpuspeed off
chkconfig cpuspeed off
service cpuspeed stop
Reboot the server


But I have the same problem every time the server is synchronizing the time!

Regards

[TheWitness]

Sorry, your log's do not lie. Maybe you should post more of your log.

TheWitness