Problems after enabling weathermap

[pvalck]

Hey,

I just installed a new cacti server with weathermap and i'm having a problem with performance.

In the past we had a jumpbox cacti server which was running ok but was getting a bit small so we installed cacti on a huge VM with RedHat. On the JumpBox we were having 11 maps without any problems.

On the new server I only have one map and as soon as this is running al my RRD-graphs are getting gaps in it.

The specs of the VM are much better then the old JumpBox.

Technical Support
General Information
Date Sat, 02 Aug 2014 00:32:55 +0200
Cacti Version 0.8.8b
Cacti OS unix
SNMP Version NET-SNMP version: 5.5
RRDTool Version RRDTool 1.3.x
Hosts 157
Graphs 3094
Data Sources Script/Command: 36
SNMP: 663
SNMP Query: 2639
Script Query: 8
Total: 3346
Poller Information
Interval 60
Type SPINE 0.8.8b Copyright 2002-2013 by The Cacti Group
Items Action[0]: 5973
Action[1]: 52
Total: 6025
Concurrent Processes 2
Max Threads 10
PHP Servers 1
Script Timeout 25
Max OID 20
Last Run Statistics Time:2.6829 Method:spine Processes:2 Threads:10 Hosts:152 HostsPerProcess:76 DataSources:5950 RRDsProcessed:3274
PHP Information
PHP Version 5.3.3
PHP OS Linux
PHP uname Linux dppucacp01.server.lan 2.6.32-431.3.1.el6.x86_64 #1 SMP Fri Dec 13 06:58:20 EST 2013 x86_64
PHP SNMP Installed
max_execution_time 300
memory_limit 512M

My questions is now, why is this and can you help?

Best regards,
Peter

[Howie]

Tell us about the maps... how many targets in total? How big (pixels) are the maps?

So for example, what is the output from...

Code: Select all

grep TARGET configs/* | wc-l
grep ^NODE configs/* | wc-l
grep ^LINK configs/* | wc-l
grep VIA configs/* | wc-l
grep 'SYSTEM STATS' ../../log/cacti.log | tail -1
grep 'WEATHERMAP STATS' ../../log/cacti.log | tail -1

?

(plus the map sizes)

[pvalck]

Howie,

Hereby the requested info:

The maps are always 1920*1080

On the new server:
TARGET = 11
NODE = 43
LINK = 16
VIA = 3
grep 'SYSTEM STATS' ../../log/cacti.log | tail -1
08/04/2014 05:01:03 PM - SYSTEM STATS: Time:2.7583 Method:spine Processes:2 Threads:10 Hosts:152 HostsPerProcess:76 DataSources:5950 RRDsProcessed:3301

grep 'WEATHERMAP STATS' ../../log/cacti.log | tail -1 didn't work here.
So i used grep 'WEATHERMAP' cacti.log | tail -1
08/04/2014 05:11:05 PM - WEATHERMAP: Poller[0] DEBUG: [weathermap_memory_check@poller-common.php:11] MEM Final: memory_get_usage() says 11.9MBytes used. Limit is 512M

On the old server:
TARGET = 1005
NODE = 1758
LINK = 341
VIA = 39

grep 'SYSTEM STATS' ../../log/cacti.log | tail -1
SYSTEM STATS: Time:20.5048 Method:cmd.php Processes:5 Threads:N/A Hosts:254 HostsPerProcess:51 DataSources:6241 RRDsProcessed:4806

Below you can find a screenshot of the configs we normally run and their sizes:


I'm at the moment migrating systems to the new cacti so not everything is yet in the new system.

Bets regards,
Peter

[pvalck]

Howie,

I found the cause of my problem and it wasn't Weathermap.

Probably Weathermap triggered it because of the extra load on the system.

The problem was that i had 2 processes and 10 threads defined in Spine. By lowering the amount of threads to 5 and changing the amount of processes to 4 all the gaps are gone and Weathermap is running without a problem.

Best regards,
Peter

[Howie]

Cool, because your maps didn't seem too heavy...