Cacti is a complete network graphing solution designed to harness the power of RRDTool’s data storage and graphing functionality. Cacti provides a fast poller, advanced graph templating, multiple data acquisition methods, and user management features out of the box. All of this is wrapped in an intuitive, easy to use interface that makes sense for LAN-sized installations up to complex networks with hundreds of devices.
After trying Cacti for a few days, I found it very useful, and full of potential, but I also found that it still doesn’t provide many useful templates.
In my case I tried to graph the usual performance metrics of an HP-UX box, and I couldn’t get it! So I developped some Data Queries, Data Templates and Graph Templates, and unified them in a Host Template.
For the moment I’ve only made an HP-UX template, but I plan to do a Linux host template basing me on the same (or corresponding) metrics.
The HP-UX Host template provides the following graphs:
* *CPU Usage*: Sys, Nice, User and Idle CPU time
* Jobs Average: 1, 5, and 15min Jobs Average
* Memory: Total vs. Used/Free
* Swap: Total vs. Used/Free
* Processes: Number of running processes
* Users: Number of logged in users
* Lan traffic (associated to Data Query “HPUX Lan”): In/Out Traffic and In/Out Errors
* Filesystem (associated to Data Query “HPUX Filesystems”): Total vs Used/Free blocks
Feel free to download the template, and use and/or redistribute it under the terms of the GPL license.
Bonjour,
J’utilise ton template, malheuresment le resultat est erroné, Cpu load a 100,140,200 par ex, ou bien espace disque exprime en Mo au lieu de Go.
Pourrais tu m’aider a résoudre ce problème ?
D’avance merci,
Clement
Hi Clement,
What Cacti graphs is the result of query to the HP-UX snmp agent. Actually I think the results are correct. Let me explain a bit:
Regarding “Jobs Averages (1/5/15min)”, the HP-UNIX mib says that it is the average number of jobs in the last 1/5/15 minute(s) * 100. That is why the graph shows so high values.
Regarding filesytem usage, the result is expressed in Blocks (not MB, neither GB). So when the graph says “1 M”, it means 1 “Mega” _blocks_, ie. ~1000 Blocks. If that filesystem blocksize is 1k, it would mean 1 GiB. If it were a 8k block filesystem, you would have to multiply by 8k. Anyway, the Total size and the available size are expressed in the same unit, so, IMHO, the graphs are corrects and useful like this.
I hope it solves your doubts about the template, and that you can find it useful. Thank you for testing it, and for your feedback.
Hi Alex,
Thank your template
…
(I am sorry, but I have bad English^^)
…
I found your HP-UX cacti template. I’m very happy^^
but I think the template have a mistake..
That is a disk free graphing.
You say, filesystem results is expressed in blocks.
but when the filesystem oversize 1G,
I cannot understand my server graph.
filesystem totalsize(byte) block(cacti graph)
/aaa 3072000 3.07M
/bbb 20930560 5.23M
Hi namgon,
Please do a df -g /bbb and look at the value of fragment size. In your case I suppose it will be 4096, meaning you have a 4k block size filesystem. fstyp -v would give you the same result, but need to be run as root, which is not necesarry for df:
adumont@kayak$ df -g /alex
/alex (/dev/vx/dsk/rootdg/volalex) :
8192 file system block size 1024 fragment size
2096128 total blocks 1583762 total free blocks
1484847 allocated free blocks 395948 total i-nodes
395939 total free i-nodes 395939 allocated free i-nodes
16777227 file system id vxfs file system type
0 flags 255 file system name length
/alex file system specific string
Remember that the graph in Cacti shows the *numbers of blocks*, which is not the same as the number of Bytes (there is the block size factor multiplying them)
Bytes number = Fragment size x Blocks number
Note: In df -g output (respectively fstyp -v), don’t get confused by “file system block size” (resp. f_bsize) . You have to look at the “fragment size” (resp. f_frsize). (See http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=107147)
Please tell me if it explains what you see.
Regards,
Alex
Hi, Alex.
Thank you.~
I solved the problem. you are right!!!!
Have a nice day~~~~~~~!!!
Merci Alex !!
Apres quelques changements (blocks vers octets) ton template fonctionne a merveilles !