Gaps in Graphs - VmWare

[Keene44]

Hello All,

I am in a constant battle to remove gaps from graphs. It is happening on approximately all my devices. I am seeing on occasion SNMP Timeouts on my Netware servers but nothing other than that. Here are my statistics:

12/12/2006 10:57:40 AM - SYSTEM STATS: Time:157.7842 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:5666
12/12/2006 10:52:50 AM - SYSTEM STATS: Time:167.7692 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:5666
12/12/2006 10:47:42 AM - SYSTEM STATS: Time:159.7320 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:5666
12/12/2006 10:44:26 AM - RECACHE STATS: RecacheTime:11.7722 HostsRecached:2
12/12/2006 10:44:13 AM - SYSTEM STATS: Time:251.1316 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:5615
12/12/2006 10:35:42 AM - SYSTEM STATS: Time:39.5612 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:245
12/12/2006 10:35:37 AM - SYSTEM STATS: Time:335.2644 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:5666
12/12/2006 10:27:50 AM - SYSTEM STATS: Time:167.7390 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:5659
12/12/2006 10:22:48 AM - SYSTEM STATS: Time:166.8496 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:5666
12/12/2006 10:17:49 AM - RECACHE STATS: RecacheTime:11.7848 HostsRecached:1
12/12/2006 10:17:36 AM - SYSTEM STATS: Time:154.5522 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:5661
12/12/2006 10:12:53 AM - SYSTEM STATS: Time:102.6102 Method:cactid Processes:1 Threads:25 Hosts:92 HostsPerProcess:92 DataSources:23884 RRDsProcessed:5666

I have attached some images of the graphs and my Poller settings.

[Keene44]

Forgot to mention my server info:

OS: Redhat ES 4 on VmWare ESX
MEM: 1GB

[gandalf]

This may be due to a hanging script. You may try to use 2 cactid processes to split the load. Using about 1-2 times the number of CPU/Cores is a good starting point
Reinhard

[Keene44]

I tried increasing the number of processes to 2 and gaps in the graphs unfortunately did not improve. Here are my stats:

12/14/2006 11:11:27 AM - SYSTEM STATS: Time:384.8799 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 11:08:24 AM - SYSTEM STATS: Time:502.1625 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:59:58 AM - SYSTEM STATS: Time:295.8142 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:54:16 AM - SYSTEM STATS: Time:253.5080 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:49:43 AM - SYSTEM STATS: Time:281.0801 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:44:20 AM - SYSTEM STATS: Time:258.1517 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:39:36 AM - SYSTEM STATS: Time:274.5633 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:36:07 AM - SYSTEM STATS: Time:365.4800 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:29:13 AM - SYSTEM STATS: Time:251.1822 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5703
12/14/2006 10:24:42 AM - SYSTEM STATS: Time:277.1070 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:19:34 AM - SYSTEM STATS: Time:272.2467 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:14:10 AM - SYSTEM STATS: Time:218.2663 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:09:41 AM - SYSTEM STATS: Time:279.3703 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 10:06:50 AM - SYSTEM STATS: Time:407.5257 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 09:59:18 AM - SYSTEM STATS: Time:256.1800 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 09:54:55 AM - SYSTEM STATS: Time:293.2848 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727
12/14/2006 09:49:08 AM - SYSTEM STATS: Time:246.0392 Method:cactid Processes:2 Threads:20 Hosts:95 HostsPerProcess:48 DataSources:23999 RRDsProcessed:5727

Below is a picture of a graph with gaps in it.

[gandalf]

Did you made any progress trying to indentify a "slow script"?
Reinhard

[Keene44]

I'm not really sure how to check for a "slow script". Any tips on how to look for such a thing?

Thanks

[Keene44]

Looking in apache log reveals following:

libpng error: Write Error
libpng error: Write Error
libpng error: Write Error

[gandalf]

Looking in apache log reveals following:

libpng error: Write Error
libpng error: Write Error
libpng error: Write Error

These are from Graph Creation only. That's not the clue
Reinhard

PS: There's a poller performnce plugin out there. I wrote the basics but have not used it lately. It's a bit ugly, yes, but you may try it
Reinhard

[scooter1966]

I also run cacti in a VM in ESX server. Fedora had no chance of maintaining an accurate time in a VM. I set a chron job to restart the NTPD to sync time 24 times a day at first. Numerous gaps in all graphs! After much trial and error I settled on a frequency of 2 minutes. That's right. Two minutes! Sounds extreme but no more gaps!