BigIP v9 Template

[clandagent]

Hi I tried your template but when i want to create the graphs i get
this message for the virtual server:

+ Running data query [14].
+ Found type = '3' [snmp query].
+ Found data query XML file at '/var/www/tpadmin.thinprint.de/cacti/resource/snmp_queries/f5_bigip_vs.xml'
+ XML file parsed ok.
+ Executing SNMP walk for list of indexes @ '.1.3.6.1.4.1.3375.2.2.10.2.3.1.1'
+ No SNMP data returned
+ Found data query XML file at '/var/www/tpadmin.thinprint.de/cacti/resource/snmp_queries/f5_bigip_vs.xml'
+ Found data query XML file at '/var/www/tpadmin.thinprint.de/cacti/resource/snmp_queries/f5_bigip_vs.xml'
+ Found data query XML file at '/var/www/tpadmin.thinprint.de/cacti/resource/snmp_queries/f5_bigip_vs.xml'

When i use snmpwalk i get the names for my virtual hosts.
I read something about "UCD-SNMP 4.x (I'm using 5.x)" but can't find
a sollution.

How can i fix it?

Thanks Hans

I get that exact same result. Have you found a solution?

[MyWebsiteAdviser]

Hi,

Thanks a lot for the template Works perfect.

[aleu]

Is there a way to monitor the CPU utilization of the BigIP? SNMP-NET based CPU templates do not show correct information when the appliance is running version 9.3 or later. None of the Linux tools shows the correct information, because 100% CPU time is reserved (taken) by TMM. Is there a way to monitor how much of the CPU time taken by TMM is really used in order to balance traffic? Some SNMP OID perhaps?

[pwohlers]

The calculation is kind of a pain. Here's a pic of the f5 doc explaining how to do it. I imagine this can be implemented as a php script, but haven't gotten that far yet :-/

In the meantime, tmstat from the command line will give you some idea of load.

[pwohlers]

Version 0.8.6j

When trying to update my bigip host template to:
F5_BigIP_v9-Host_Template-rev_6.xml

I get the following error when trying to import it:

Error: XML: Cacti version does not exist.

I had seen something in the forums that this might happen if the template was created with a newer version of cacti, but in this case, I thought newer versions of cacti were OK with older templates.

Any suggestions?

TIA

[aleu]

The calculation is kind of a pain. Here's a pic of the f5 doc explaining how to do it. I imagine this can be implemented as a php script, but haven't gotten that far yet :-/

In the meantime, tmstat from the command line will give you some idea of load.

pwohlers,

Thanks for your response. Yes, I this seems to be a bit of pain. Well, I hope that someone more advanced in Cacti will help. Anyone?

Thanks,
AL

[pwohlers]

Version 0.8.6j

When trying to update my bigip host template to:
F5_BigIP_v9-Host_Template-rev_6.xml

I get the following error when trying to import it:

Error: XML: Cacti version does not exist.

I had seen something in the forums that this might happen if the template was created with a newer version of cacti, but in this case, I thought newer versions of cacti were OK with older templates.

Any suggestions?

TIA

self-help answer:

this fixed it:

in file:
/var/www/cacti/include/config_arrays.php


$hash_version_codes = array(
"0.8.4" => "0000",
"0.8.5" => "0001",
"0.8.5a" => "0002",
"0.8.6" => "0003",
"0.8.6a" => "0004",
"0.8.6b" => "0005",
"0.8.6c" => "0006",
"0.8.6d" => "0007",
"0.8.6e" => "0008",
"0.8.6f" => "0009",
"0.8.6g" => "0010",
"0.8.6h" => "0011",
"0.8.6i" => "0012",
+ "0.8.6j" => "0013"
);

adding that one line fixed it. Why that wasn't automagically there i'll attribute to some sort of pilot error on the upgrade/migration we did a few months back. ymmv.

[khufure]

Works great. One gotcha is to be sure to use v2c and configure your bigip appropriately. v1 does not work!

[aboling]

Sorry to cross-post like a jerk, but I ran into some funny issues importing this into a cacti setup I've "inherited" where I work.

The "VS connections" related queries were returning the infamous "duplicate data source" error because the "F5 BigIP - VS Traffic" data template was importing with duplicate "traffic_in" data source items. If you're running into the same problem, it's probably related to having extra interface data templates. Below is the thread where I'm trying to get it straightened out:

http://forums.cacti.net/viewtopic.php?t=24232


I'm dropping a link here in case someone has some insight into why this might be happening, if it is not in fact related to having extra interface data templates.

[citizen_elah]

I am using the attached perl script, data template, and graph template to monitor TMM CPU Utilization. The perl script goes in <cacti_path>/scripts/

[orangepeelbeef]

That CPU template is not correct. First off, you need to pull twice waiting 10 seconds between and do the calculations on the delta's. Secondly, the TMM is nearly always 100% by design, so there isn't much point to graphing that out.

[citizen_elah]

The calculation is verbatim out of F5's documentation for TMM's utilization of the CPU. It works quite well in my environment, matching the F5 performance graph line for TMM exactly.

[aleu]

The calculation is verbatim out of F5's documentation for TMM's utilization of the CPU. It works quite well in my environment, matching the F5 performance graph line for TMM exactly.

Agreed, but it does not look that your perl script waits for 10 seconds before polling the second value... F5 documentation states that you need to wait these 10 seconds.

[citizen_elah]

Missed that. Doesn't seem to impact my results, however. I suppose that 10s delta versus the 5 min delta I already have is to smooth the data? In any event, if you want to edit the script, you can add a second poll and change the calculation:

Code: Select all

sleep 10;
my $polled_oids_1 = $session->get_request(
                  -varbindlist =>
                  [$tmmTotalCyl, $tmmIdleCyl, $tmmSleepCyl] );

my $tmm_cpu = (( ($polled_oids_1->{$tmmTotalCyl} - $polled_oids_0->{$tmmTotalCyl}) -
                ( ($polled_oids_1->{$tmmIdleCyl} - $polled_oids_0->{$tmmIdleCyl}) +
                ($polled_oids_1->{$tmmSleepCyl} - $polled_oids_0->{$tmmSleepCyl}) ))
                / ($polled_oids_1->{$tmmTotalCyl} - $polled_oids_0->{$tmmTotalCyl}) ) * 100 ;

[/code]

[aleu]

Missed that. Doesn't seem to impact my results, however. I suppose that 10s delta versus the 5 min delta I already have is to smooth the data? In any event, if you want to the script, you can add a second poll and change the calculation...

Thanks for your contribution!