Poller (cmd.php or cactid) processes instances in cacti.log

[twelzy]

Hi everyone!

I think that whould be great if we have the poller (cmd.php or cactid) processes instances/numbers in cacti.log.

Like: Poller[#]...

Let me try to explain better...

I'm running cactid but my poller time is always around 296 seconds.

12/06/2005 09:24:57 PM - SYSTEM STATS: Time:296.5770 Method:cactid Processes:20 Threads:40 Hosts:141 HostsPerProcess:8 DataSources:1245 RRDsProcessed:535

So I take a look at my cacti.log file.
(My cacti.log huge!)
Then I discovered that I'm having timeout problems with some data queries.

12/06/2005 09:20:11 PM - CACTID: Poller[0] Time: 10.2918 s, Threads: 40, Hosts: 9
12/06/2005 09:20:16 PM - CACTID: Poller[0] Time: 15.2870 s, Threads: 40, Hosts: 9
12/06/2005 09:20:16 PM - CACTID: Poller[0] Time: 15.2982 s, Threads: 40, Hosts: 9
12/06/2005 09:20:17 PM - CACTID: Poller[0] Time: 16.8390 s, Threads: 40, Hosts: 9
12/06/2005 09:20:19 PM - CACTID: Poller[0] Time: 18.4560 s, Threads: 40, Hosts: 9
12/06/2005 09:20:20 PM - CACTID: Poller[0] Time: 19.5749 s, Threads: 40, Hosts: 9
12/06/2005 09:20:27 PM - CACTID: Poller[0] Time: 26.3503 s, Threads: 40, Hosts: 9
12/06/2005 09:20:29 PM - CACTID: Poller[0] Time: 28.3480 s, Threads: 40, Hosts: 9
12/06/2005 09:20:30 PM - CACTID: Poller[0] Time: 29.4862 s, Threads: 40, Hosts: 9
12/06/2005 09:20:43 PM - CACTID: Poller[0] Time: 42.3970 s, Threads: 40, Hosts: 9
12/06/2005 09:20:50 PM - CACTID: Poller[0] Time: 49.3494 s, Threads: 40, Hosts: 6
12/06/2005 09:20:50 PM - CACTID: Poller[0] Time: 49.4570 s, Threads: 40, Hosts: 9
12/06/2005 09:21:16 PM - CACTID: Poller[0] Time: 75.0350 s, Threads: 40, Hosts: 9
12/06/2005 09:21:27 PM - CACTID: Poller[0] Time: 86.3770 s, Threads: 40, Hosts: 9
12/06/2005 09:21:40 PM - CACTID: Poller[0] Time: 99.4987 s, Threads: 40, Hosts: 9
12/06/2005 09:21:43 PM - CACTID: Poller[0] Time: 101.9637 s, Threads: 40, Hosts: 9
12/06/2005 09:22:59 PM - CACTID: Poller[0] Time: 177.2682 s, Threads: 40, Hosts: 8
12/06/2005 09:24:56 PM - CACTID: Poller[0] ERROR: Cactid Timed Out While Processing Hosts Internal
12/06/2005 09:24:56 PM - CACTID: Poller[0] Time: 295.4492 s, Threads: 40, Hosts: 9

If I have the information of what specific poller process instance (for example: Poller[3]) is having this timeout problems, I could search the cacti.log looking for this specific instance only.

Thanks in advance!
(...and sorry for my poor English!)

[TheWitness]

You would be much better served by reducing processes to no more than 2 times the number of processes and threads closer to 15-20. Give it a try.

I have close to 15k data sources and 6k rrd files and poll with 4 processes, 30 threads in less than 40 seconds.

TheWitness