I’ve released a new version of my Cacti host templates for Unix systems (HP-UX, Solaris and Linux).
This new version introduces the following changes:
* ucd Load Average GT seemed somehow messed up in v1.2 (load5min missing in xml, reported by eyechart). I think it is solved now.
* New Uptime GT (all OS, contributed by Mikf here)
*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 my Cacti host templates for Unix, and use and/or redistribute it under the terms of the GPL license.
You may also check this post on the Cacti forum, for further information and/or post comments here.
Has the templates been tested with cacti 0.8.6h?
Yes, some people use them successfully on 0.8.6h. Check this post.
Hi-
Curious on you your CPU Utiliz and Load, work on a multi-proc system. Is this the average across the server?
@wcskid:
The CPU and Load average are average metrics on the system. It is not the usage/load of a specific processor.
I quote here the Net-SNMP FAQ:
[...]“the CPU statistics (both original percentages, and the newer raw statistics) both refer to the system as a whole.”[...]
Thanks for the info. BTW I just cannot get the HP 80 Lan.. or HP 90 File.. to load. I’ve tried importing the templates several times multiple ways. Using Cacti 8.6.h on a W2K server. They show up under the graph templates but, nothing shows up to “graph” under the device or when you “create a new graph”.
Suggestions, workarounds?
@WCSKid:
You have to manually copy hpux_fs.xml and hpux_lan.xml files into you Cacti installation dir. Those are needed so that both queries (hpux lan and fs) could work correctly.
Have a look at the INSTALL file shipped within the tarball.
Thanks, did that already, I don’t want to fill your blog should I continue at cacti.org? the fs and lan come up under the device and say they are working fine.
But can’t create a new graph for fs or LAN. from either “New graphs” or “Devices” “add a graph”.
Please, try importing again the cacti_host_template__hpux_host-1.3.xml file (overwritting the already imported templates. Do not remove any templates before reimporting them). Sometimes it solves the problem (don’t know why!?).
Then recreate the device in Cacti, or at least rerun the Data Queries (click the green circle).
If it doesn’t work, I think that you may try to post to Cacti’s forum, so more users can help you.
Found the issue, my 8.6h was not patched thier patches fix the issues.
Thanks!
Hello,
Just wanted to say thank you for the templates for Cacti. They are a real time saver and designed well.
I just wanted to point a small problem I had that took me a long while to figure out. The data template for ucd_ssCpuRawWait had a space on the end of the OID which prevented it from obtaining the wait information. Other than that, everything else works great!
Are you planning on adding any additional graphs soon? There quite a few more that I think you could add that would be very useful.
Thanks again,
John
Hello John,
Thanks a lot for your feedback and for pointing out the errata in the definition of ucd_ssCpuRawWait’s OID. Will be corrected in next release.
What graph would you like to be added? Do you have any example?
Hello –
The memory graphs for a UX 11i 9000/800 box seems to be out by a factor of 1000.
The total amount of memory being reported is 4M when it should be 4GB.
Thanks..
@chimaera: yes, the unit of the memory graph is Kb, so 1M in the graph is “1 Mega KB”, that is 1GB.
adumont-
I have a couple of HP-UX 11.x boxes I am using your host template on. For the most part they all seem to be functioning well however, on one of the boxes the load is measured by m’s (200 m, 300 m 400 m) -vs- 1.0 , 2.0, 3.0 etc.
Any idea why this would be happening?
Well, “200m” means 200 milli-units. In that case, you have to understand the load is 0,200 .
This happends because the load average reported by snmpd have to be divided by 100. By the way, I think there is noting wrong about that. It´s just that the host is not very busy, is it?
Hi,
thanks for the work you’ve done on the HP-UX templates.
In the filesystem graphs, you use blocks to display the file system usage. I need to display the usage in bytes. I can create a cdef that will give the usage in bytes for an individual graph, but I am not sure how to apply this to the whole template to allow this as an option when creating graphs for a host.
I would appreciate any help if you have time.
Thanks.
Cheers, Barrie
Hi Alex
I want to thank you for the great job!
We have a lot of HP-UX server that needs monitoring and hpux default snmp is the only (not invasive) method allowed.
Your template solved me a great number of problems, but my chiefs don’t like some things like filesystem measurement (expressed in blocks instead of bytes) and the dimension of swap and memory (expressed in kilobytes instead of bytes).
After many days of suffering (i’m new to cacti/snmp/php/xml, gosh!), i solved all the problems adding 6 new cdef, adding an snmp entry in your file hpux_fs.xml and modifying your hp-ux host template
I’ve also rearranged some colours and stack to have a immediate impression on what’s going wrong… now if i see an all red graph, i’m sure that there is a problem :o))).
Now my chiefs are very happy!
I’ve exported the modified hpux template and i’ll mail you soon with all the steps i’ve done to achieve this result.
Can you take a look at my files to see if it’s possible to submit this changes in your next release?
I think that they will be very useful.
Thanks.
Diego
sorry for the spam, my email address in last post was wrong.
sorry also for my bad english :D
Hi Diego,
Thanks for your comments, your detailled mail about the
modifications you’ve made to the template, and also for the template zip file. (The detail level of your mail is quite impressive!).
I’m looking forward to trying it (when I get to find some free time…).
Best Regards,
Alex
Hi Alex,
I would like to try out Diego’s templates if possible. It sounds like Diego has done what I am trying to do (and more). Could you send me the templates if possible, or maybe send my email address to Diego so he could send them to me.
I would like to understand the process Diego used to modify the host template so that I can do that in the future as well.
Thanks for the great work you’ve done.
Cheers, Barrie
Hi Alex,
I haven’t heard from Diego as yet, would it be possible for you to send me his templates? I am currently implementing the monitoring of our HP Superdomes using Cacti. The modifications he has described would be a lot of use to me.
Thanks.
Cheers, Barrie
I need a good working example of /usr/local/share/snmp/snmpd.conf for a Solaris 9 SPARC installation. I have installed the latest Net-SNMP from source and have run the basic config – but I have problem understanding all the weird questions!
I need to monitor CPU, Memory, Disk utilisation, file size changes, etc – can someone please help.
Is there any chance of getting the graphs modified or new ones added that actually shows Use% which indicates how much space is used? The total/used graph makes it difficult to implement threshold monitoring of say 90% across all filesystems for alerting purposes.
Please note that this only relates to the mounted filesystems on the host.
Your templates are great and nearly everything works. I get CPU and lan information but I get no informations regarding filesystems.
Do you have any idea?
Hello,
Great Job!
However, I have one small problem. The ucd 80 lan template does not seem to work on Gigabit interfaces. It shows a maximum of just slightly over 100Mb but the server itselft is showing 260+Mb. Have you seen this before?
Thanks,
John
Hello !!
The templates for Linux works well, good job. But I have still problems with snmp.conf please can someone send me conf file. I have still problems with some graphs, I thing that the problem is in the snmp configuration on the Linux machine.
Thanks
Papik
Hi Alex,
Just an update, we are using the HP-UX templates in production now. All seems to be working well.
Cheers, Barrie
Hi,
I am just wondering why the linux templates (memory, detailed memory, swap) show memory in kilobytes and therefore the wrong amount ? i.e 512mb ram is shown as 512k ? 1 gig swap is shown as 1M ?
thanks
Kashif
Nice work.
I would suggest a modification of the cpu graphs so it also shows IOwait. Using the CDEF function cdef=100,a,-,b,-,c,-,d,- one gets the IOwait and it is not needed to use a scale function anymore.
Too bad the hp-unix MIB does not have numbers you get with swapinfo -tam. The total swap with pseudo swap included and the reserved swap numbers are missing, like also the vmstat and iostat numbers.
bye, Paul
Is there a updated version of the template with the IOWait modification that Paul mentions? I can’t get the cdef applied correctly
I’m sorry, but for the moment, there is no new version of my templates. I don’t have enough free time to develope them.
I got it right by myself. Putting all unaccounted time in IOWait feels a bit shaky but it definitely better than the old graphs.. I’m comparing them with sar -u 5 1 graphs now
Greetings Alex,
Having issue with solaris templates getting filesystem information…please advise.
+ Running data query [8].
+ Found type = ‘6 ‘[script query].
+ Found data query XML file at ‘/var/www/cacti/resource/script_server/host_disk.xml’
+ XML file parsed ok.
+ Executing script for list of indexes ‘/usr/bin/php -q /var/www/cacti/scripts/ss_host_disk.php 10.240.0.44 849 1:161:500:public:::MD5::[None] index’
+ Executing script query ‘/usr/bin/php -q /var/www/cacti/scripts/ss_host_disk.php 10.240.0.44 849 1:161:500:public:::MD5::[None] query index’
+ Executing script query ‘/usr/bin/php -q /var/www/cacti/scripts/ss_host_disk.php 10.240.0.44 849 1:161:500:public:::MD5::[None] query description’
+ Executing script query ‘/usr/bin/php -q /var/www/cacti/scripts/ss_host_disk.php 10.240.0.44 849 1:161:500:public:::MD5::[None] query sau’
+ Found data query XML file at ‘/var/www/cacti/resource/script_server/host_disk.xml’
+ Found data query XML file at ‘/var/www/cacti/resource/script_server/host_disk.xml’
+ Found data query XML file at ‘/var/www/cacti/resource/script_server/host_disk.xml’
Disregard my previous post, unix admin verified that snmp oid for fs on solaris isn’t there by default.
@Kelly: Are you running NetSNMP for Solaris?
If i remember well, the standard Solaris SNMPd won’t report enough information (like FS,…). You have to use NetSNMP for Solaris.
Thanks for you templates. It made things so much easier to collect data from SUN and Linux.
Maybe you could give me some pointer about collecting running process information. I see template for W32 but not for Linux.
So far I found out that I have to use snmpwalk and query the information from hrSWRun and hrSWRunperf
I have problems with your template and thold plugin, please can you help-me ? See the link http://forums.cacti.net/viewtopic.php?t=20224
lucianogomes_rj@Hotmail.com
Hi Luciano,
I’ve not tryed thold, but anyway, if you look at this link
http://forums.cacti.net//files/thold_functionsphp_135.patch, you’ll see the following piece of code from thold:
+ if ($cdef['type'] == 4) {
+ $cdef['type'] = 6;
+ switch ($cdef['value']) {
+ case ‘CURRENT_DATA_SOURCE’:
+ $cdef['value'] = get_current_value($rra, $ds, 0);
+ break;
+ case ‘CURRENT_GRAPH_MAXIMUM_VALUE’:
+ $cdef['value'] = get_current_value($rra, ‘upper_limit’, 0);
+ break;
+ case ‘CURRENT_GRAPH_MINIMUM_VALUE’:
+ $cdef['value'] = get_current_value($rra, ‘lower_limit’, 0);
+ break;
+ case ‘CURRENT_DS_MINIMUM_VALUE’:
+ $cdef['value'] = get_current_value($rra, ‘rrd_minimum’, 0);
+ break;
+ case ‘CURRENT_DS_MAXIMUM_VALUE’:
+ $cdef['value'] = get_current_value($rra, ‘rrd_maximum’, 0);
+ break;
+ default:
+ print “CDEF property not implemented yet: ” . $cdef['value'];
+ break;
+ }
As you can see, the implemented CDEF are:
CURRENT_DATA_SOURCE
CURRENT_GRAPH_MAXIMUM_VALUE
CURRENT_GRAPH_MINIMUM_VALUE
…
IMHO, it lacks a “CURRENT_GRAPH_VALUE” CDEF…
The problem seems to come from thold not having implemented the CDEF.
The graph already include the correct value you look for, so try to get thold get to use this value directly.
Alex;
Just wanted to thank you for posting the templates. I’m having a little trouble, I just posted in your original thread on the cacti forums, if you have a chance to look through my post and help, I’d appreciate it greatly. To say I’m a bit out of my league troubleshooting this is somewhat of an understatement.
Sincerely,
Ron
Hello i use your template but i have some probleme do you please help me
http://forums.cacti.net/viewtopic.php?t=11634&postdays=0&postorder=asc&start=30
I have problem i instal like the reamde says:
and i have this message:
[quote]
Running data query [10].
+ Found type = ‘3′ [snmp query].
+ Found data query XML file at ‘/var/www/html/cacti/resource/snmp_queries/hpux_fs.xml’
+ Error parsing XML file into an array.
+ Found data query XML file at ‘/var/www/html/cacti/resource/snmp_queries/hpux_fs.xml’
+ Found data query XML file at ‘/var/www/html/cacti/resource/snmp_queries/hpux_fs.xml’
+ Found data query XML file at ‘/var/www/html/cacti/resource/snmp_queries/hpux_fs.xml’
[/quote]
please someone can help me please
viva el idioma de Cervantes¡¡¡
He de decir que , aunque suene a peloteo, las graficas del script-man de Alex solucionan muchos de los problemas tipicos de…’es que va muy lento el servidor’, ‘es que los usuarios dicen que va muy despacio…’, de un solo vistazo ves como esta el sistema y si algo no te cuadra siempre puedes entrar al sistema a comprobarlo tu mismo.
Cuando tenga tiempi i y ganas intentare tunear mi cacti un poco aunque con el curro de Alex no se si queda algo por tunear ya
gracias por tu afan de superacion, ;-P
hi, i have some troubles with disk usage graphs
i got /var/spool/imap1 /var/spool/imap2 .. /var/spool/imap13 partitions
in graph title i got only /var/spool/imap for every partition
Add disk i/o support ;)