UNIX Template Set for Cacti: HP-UX, Solaris, Linux

mjjzr · Joined: 10 Oct 2005 Posts: 8

Hi adumont
Thanks for your work, that's save me a lot of time.

I've imported to a 0.86h and looks ok.

eyechart · Joined: 25 Apr 2002 Posts: 49

the 5min load average seems messed up. It doesn't use SNMP, instead it uses the perl script. Since I am monitoring remote linux/unix systems from a windows box, this doesn't seem to work.

Can the templates be altered so that the 5min average uses SNMP? thanks.

-ec

adumont · Joined: 26 Nov 2005 Posts: 16 Location: Spain

For HP-UX template, I use the following DS:

hpux_AvgJobs5: Get SNMP Data
OID=.1.3.6.1.4.1.11.2.3.1.1.4.0

For Sun and Linux templates, I use the following DS:

ucd/net - Load Average - 5 Minute: Get SNMP Data
OID=.1.3.6.1.4.1.2021.10.1.3.2

Both DS are of type SNMP Data.

You may be refering to "Unix - Load Average" DS, but I don't use it in my Hosts Templates.

eyechart · Joined: 25 Apr 2002 Posts: 49

I don't see that OID specified in the latest XML for linux. can you double check this?

I see the OIDs for 1 minuted and 15 minute, just not 5 minute.

-ec

adumont · Joined: 26 Nov 2005 Posts: 16 Location: Spain

@ec: You are right.

However, I don't know where is the problem. In my cacti installation the Graph Template "ucd 01 load average" is correctly defined (using DS load_5min).

When I export the full Linux host template , and I import it again in another "empty" Cacti installation, I loose the load_5min DS, and the GT is no correctly defined. I also loose the data query attached to the Host Template (the DQ are imported, but in the HT definition no DQ appear)!

I don't know where the problem is. Is it in the export process? The import process? Maybe it is also in my installation Linux.

I can share my Full cacti DB dump if someone wants to try exporting and importing the templates again, to see if the problem is reproduced.

Importing the full DB shows that the Host Template is correctly defined. (But it's not an easy way to share the templates between people).

adumont · Joined: 26 Nov 2005 Posts: 16 Location: Spain

I've released a new version my Cacti host templates for Unix systems (HP-UX, Solaris and Linux).

This new version introduces the following changes:
* ucd Load Average GT seemed messed up (load5min missing, reported by eyechart). I think it is solved now.
* New Uptime GT (all OS, contributed by Mikf http://forums.cacti.net/viewtopic.php?t=5951)

WARNING: Import each Host Template twice!!
It seems that sometimes the templates are not correctly imported after the first import (associated grath templates and/or data queries are missing). Importing the templates *once again* seems to solve the problem (at least in cacti 0.8.6g).

Feel free to download version 1.3 of the Unix Template Set for Cacti, and use and/or redistribute it under the terms of the GPL license.

You may also check this link and leave a comment there if you use and like these templates:
http://adumont.hopto.org/~adumont/blog/tag/cacti .

For ec: Thanks for reporting the problem with load5min. I hope it is solved in this release. At least in my instalation it seems to be fine. Anyway, have a look at the warning.

eyechart · Joined: 25 Apr 2002 Posts: 49

ok, the 1.3 release fixed the load average graphs. I had to delete the current graphs and RRAs that I had and create new ones based off of the new template for it to work though.

I like the uptime graph too, nice addition.

Thanks for fixing this. This is really a great set of graphs.

-ec

laffen · Cacti User Joined: 13 Jul 2004 Posts: 52 Location: Norway

Have these templates been tested with 0.8.6h?

eyechart · Joined: 25 Apr 2002 Posts: 49

I am running on .86h and 1.3 fixed the only problems I was having.

-ec

Antoine Musso · Joined: 12 Dec 2005 Posts: 12

Can you please edit your first post so it get the lastest version ? :o) Thanks!

wintech2003 · Joined: 17 May 2005 Posts: 12

I have a problem with:

ucd 10 Memory
ucd 11 Swap
ucd 12 Detailled Memory

which all report in Mbytes instead of Gbytes...

So on all my Linux servers my graphs say I have Total RAM 1MB or 2MB although it should say 1 or 2 GB.

Check screenshot:

This happens on all my servers (others have Slackware, others CentOS).

Can you help me fix this?

adumont · Joined: 26 Nov 2005 Posts: 16 Location: Spain

Hi,

When the graph show "1M" of unit KB (not B), this means: 1 million (or 1024x1024) KB, that is 1GB, not 1MB.

Regards,
Alex

wintech2003 · Joined: 17 May 2005 Posts: 12

lol you're right I didn't noticed it.

Is there a way to make it report bytes instead of kilobytes so that it says GB?

guidob · Joined: 16 Jun 2006 Posts: 1

I have exactly the same problem. Running on CentOS, Slackware, RedHat, SuSe & Fedora.

lrhazi · Joined: 26 Jun 2006 Posts: 4

Any idea why my lan usage graphs never show any data?
I installed and used the above templates for several servers and added most of the defined graphs... added graphs for lan, selected eth0 1 and 2 from the discovered interfaces... but no data ever shows up....
all the pther graphs, mem usage, cpu disk usage... work fine.

How do I debug this? If this is not related to these specific templates, please move this post whereeve it belongs.

Thanks a lot.
Mohamed~