The templates used to manage custom scripts are much simpler than the 
templates for custom SNMP MIBs, but using SNMP will reduce the 
maintenance in the long run.  By executing scripts from SNMP you can 
manage all your data collection through SNMP, and then distribute your 
config and scripts to all the necessary servers, eliminating the need 
for ssh keys and user accounts on end systems.  Templates will get you 
the extra granularity you would like in the graphs as well.

There are two shortcomings I have found with Cacti:

1) It is unable to manage data with a varying number if items to be 
plotted.  For example if you were trying to get a count of each type of 
virus received in email throughout the day, it wouldn't work.  This is 
because Cacti needs to be told what to plot in each graph, and since it 
is impossible to determine the set of viruses that will be received on 
any given day the graph can't be configured properly.  (If there is a 
way to do this, I'd love to know about it).

2) When managing a large number of hosts, or when there are a large 
number of data sources the time taken to execute the poller.php script 
can exceed 5 minutes.  This in turn leads to incorrect graphs since the 
data is not being updated at the correct intervals.  You will probably 
experience this problem with 1200 graphs.

Ronnie

Puryear Information Technology, LLC
Baton Rouge, LA * 225-706-8414
http://www.puryear-it.com

Visit http://www.puryear-it.com/publications.htm to download your free
copies of:

 "Best Practices for Managing Linux and UNIX Servers"
 "Spam Fighting and Email Security in the 21st Century"



-ray wrote:
> On Wed, 17 Jan 2007, Dustin Puryear wrote:
>
>   
>> Cacti rocks. We use it all over the place.
>>     
>
> I looked at Cacti a long time ago and it was ok.  I looked again and it 
> seems really cool.  I setup and managed to get a graph of Apache server 
> requests using the server-status handler.  It's a slight learning curve 
> dealing with all the templating.
>
> How are you guys doing data collection?  External scripts?  SNMP?  snmpd 
> calling external scripts?  I'd like a little more granularity that the 
> default stuff... for example a graph of each vmstat column would be nice. 
> Our current homegrown rrd system is about 1200 graphs and 150 data 
> collection scripts.  So the less i have to "touch" my systems (making 
> users, copying scripts, setting up crontabs, ssh keys, etc), the better.
>
> ray
>
>
> _______________________________________________
> General mailing list
> General at brlug.net
> http://mail.brlug.net/mailman/listinfo/general_brlug.net
>   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/general_brlug.net/attachments/20070120/355c6914/attachment.html 

Reply via email to