[solved] Noob questions

[b-boy]

Hi all

I am busy configuring my first Cacti install and I just wanted some feedback on if I'm taking the correct approach. I will be using Cacti to monitor several servers with our application installed.

1. Up status - Is it possible to configure the poller to test if a server is up by testing connectivity to a port , we have may clients that only allow port 80, 443 + 22 access with no ICMP.

2. Data templates - I pull all the information I require with a script that extracts data from a web page on each server ( 1min load average , 5min load average, 15min load average, root disk space, usr disk space, var disk space, inbound mail, outbound mail, mail queue....... ). Should I use 1 data template to collect all this information or is it better to split it up into the elements that you will graph together ?

3. Poller - It looks like its running if I check the logs but I don't see it collecting data

10/08/2009 11:55:02 AM - SYSTEM THOLD STATS: Time:0.0097 Tholds:0 Hosts:0
10/08/2009 11:55:02 AM - SYSTEM STATS: Time:1.2405 Method:spine Processes:1 Threads:10 Hosts:4 HostsPerProcess:4 DataSources:6 RRDsProcessed:0
10/08/2009 11:55:02 AM - SPINE: Poller[0] Time: 0.5411 s, Threads: 10, Hosts: 2
10/08/2009 11:55:01 AM - SPINE: Poller[0] Time: 0.5542 s, Threads: 10, Hosts: 3
10/08/2009 11:55:01 AM - POLLER: Poller[0] NOTE: Poller Int: '300', Cron Int: '300', Time Since Last: '300', Max Runtime '298', Poller Runs: '1'

When I run spine from the command line I get the correct info in the logs

10/08/2009 12:01:21 PM - SPINE: Poller[0] Time: 0.8223 s, Threads: 10, Hosts: 4
10/08/2009 12:01:21 PM - SPINE: Poller[0] Host[18] DS[70] SCRIPT: /usr/bin/perl /usr/share/cacti/site/scripts/medic.pl xxxxxxxxxx.com, output: Load1:0.11 Load5:0.25 Load15:0.27 Root:44 Usr:24 Var:79 MailQ:96 MailIn:14 MailOut:9
10/08/2009 12:01:21 PM - SPINE: Poller[0] Host[19] DS[73] SCRIPT: /usr/bin/perl /usr/share/cacti/site/scripts/medic.pl xxxxxxxxxxxx.com, output: U
10/08/2009 12:01:21 PM - SPINE: Poller[0] Host[19] DS[73] WARNING: Result from SCRIPT not valid. Partial Result: ...
10/08/2009 12:01:21 PM - SPINE: Poller[0] Host[19] ERROR: Empty result [xxxxxxxxxxxxx.com]: '/usr/bin/perl /usr/share/cacti/site/scripts/medic.pl xxxxxxxxxxx.com'
10/08/2009 12:01:21 PM - SPINE: Poller[0] Host[18] DS[69] SCRIPT: /usr/bin/perl /usr/share/cacti/site/scripts/medic.pl xxxxxxxxxx.com, output: Load1:0.11 Load5:0.25 Load15:0.27 Root:44 Usr:24 Var:79 MailQ:96 MailIn:14 MailOut:9
10/08/2009 12:01:20 PM - SPINE: Poller[0] Host[19] DS[72] SCRIPT: /usr/bin/perl /usr/share/cacti/site/scripts/medic.pl axxxxxxxx.com, output: U
10/08/2009 12:01:20 PM - SPINE: Poller[0] Host[19] DS[72] WARNING: Result from SCRIPT not valid. Partial Result: ...
10/08/2009 12:01:20 PM - SPINE: Poller[0] Host[19] ERROR: Empty result [xxxxxxxx.com]: '/usr/bin/perl /usr/share/cacti/site/scripts/medic.pl xxxxxxxxxxxx.com'
10/08/2009 12:01:20 PM - SPINE: Poller[0] Host[18] DS[68] SCRIPT: /usr/bin/perl /usr/share/cacti/site/scripts/medic.pl xxxxxxxxxxxx.com, output: Load1:0.11 Load5:0.25 Load15:0.27 Root:44 Usr:24 Var:79 MailQ:96 MailIn:14 MailOut:9
10/08/2009 12:01:20 PM - SPINE: Poller[0] Host[19] DS[71] SCRIPT: /usr/bin/perl /usr/share/cacti/site/scripts/medic.pl xxxxxxxxxx.com, output: U
10/08/2009 12:01:20 PM - SPINE: Poller[0] Host[19] DS[71] WARNING: Result from SCRIPT not valid. Partial Result: ...
10/08/2009 12:01:20 PM - SPINE: Poller[0] Host[19] ERROR: Empty result [xxxxxxxxxxxx.com]: '/usr/bin/perl /usr/share/cacti/site/scripts/medic.pl xxxxxxxxxxx.com'


OS Ubuntu 9.04
Cacti Version 0.8.7e
Plugin Arc 2.5
Settings 0.5
Monitor 0.8.2
Thold 0.4.1

Thanks in advance for your help
b

[b-boy]

PS. Why do i see the poller polling each of my hosts x3 ? I only have 1 data input Method, does the system poll each host multiple times for each data/graph template ?

[b-boy]

Ok, I fixed #3 , I needed to add the poller to my cron

[BSOD2600]

1) Change the Downed Device Detection to ping and select it to tcp or udp and specify the port required.

2) This becomes a trade off on what eventually of performance vs ease of use. While it might be easier to have 12 single entry data templates, which a single graph template pulls from.... there will be a performance overhead for rrdtool having to open 12 rrd files and pull the data. Additionally, it's a pain to manually set 12 items in a graph. Conversely, having a Data template with 12 items starts to reach the point of a pain to set up.

If you're going with a more automated method of setting up graphs (data queries, etc) then having all of the same/logical groupings of items in the same data template makes sense.

[b-boy]

Thanks for the feedback