[XML] Cisco IP SLA templates (AKA RTR & SAA)

[jay]

Hi

375M seems very excessive!!! I have just checked my SAA graphs for 2 years of data and it was 98K

My largest RRD file is a Cisco Memory file, with 6 months of data, which is just under 5M.

What RRA are you using for these templates?? It sounds like you are doing 1 minute polling and keeping granular data over a longer period of time.

Cheers

Jay

[osiris123d]

When I go into Admin Mode -> Data Templates -> Cisco Router SAA -> RTTMax
I see that for Associated RRA's
Hourly(1 minute average)
Hourly(1 minute average)
Daily(5 minute average)
Weekly(30 minute average)
Monthly(2 hour average)
Yearly(1 day average)

I am pretty positive that I didn't mess with these settings when I first installed the template. Is this what it should be?

Hi

375M seems very excessive!!! I have just checked my SAA graphs for 2 years of data and it was 98K

My largest RRD file is a Cisco Memory file, with 6 months of data, which is just under 5M.

What RRA are you using for these templates?? It sounds like you are doing 1 minute polling and keeping granular data over a longer period of time.

Cheers

Jay

[jay]

Hi

My RRA settings are as follows:-

Daily (5 minute average)
Weekly (30 minute average)
Monthly (2 hour average)
Yearly (1 day average)

This is the standard as far as I'm aware. I run Cacti version J. What version are you running?? Some of the later version may have additional rra settings and when you imported the template it took these settings as well. If my maths servers me well, and you have 2 hourly 1 minute average rra, then that will be a lot of granular data collected, hence the large rrd file.

If you don't want such granular data then all you need to do is to un highlight the 1 minute averages in your data template. I think this will stop you graphing at these averages, I think. Have you got a test box you could test this on first. I don't want to give you advice that could cause problems for your current graphs.

If you import any future templates I would always go into the data template and check what rra settings its using before creating graphs. I have had the same issue as you, as in it created multiple graphs at different rra settings, but luckily it was in a test enviroment and I only wanted to check the templates worked before moving into production.

Cheers

Jay

[osiris123d]

I am running Version 0.8.7a

I will play with the RRA and see what I can find. Thanks for the info.

Hi

My RRA settings are as follows:-

Daily (5 minute average)
Weekly (30 minute average)
Monthly (2 hour average)
Yearly (1 day average)

This is the standard as far as I'm aware. I run Cacti version J. What version are you running?? Some of the later version may have additional rra settings and when you imported the template it took these settings as well. If my maths servers me well, and you have 2 hourly 1 minute average rra, then that will be a lot of granular data collected, hence the large rrd file.

If you don't want such granular data then all you need to do is to un highlight the 1 minute averages in your data template. I think this will stop you graphing at these averages, I think. Have you got a test box you could test this on first. I don't want to give you advice that could cause problems for your current graphs.

If you import any future templates I would always go into the data template and check what rra settings its using before creating graphs. I have had the same issue as you, as in it created multiple graphs at different rra settings, but luckily it was in a test enviroment and I only wanted to check the templates worked before moving into production.

Cheers

Jay

[osiris123d]

Now that I think about it I might have been messing with 1 minute polling a while back. I might have created that 2nd Hourly RRA. Anyway I just deleted it and we will see what happens.

I am running Version 0.8.7a

I will play with the RRA and see what I can find. Thanks for the info.

Hi

My RRA settings are as follows:-

Daily (5 minute average)
Weekly (30 minute average)
Monthly (2 hour average)
Yearly (1 day average)

This is the standard as far as I'm aware. I run Cacti version J. What version are you running?? Some of the later version may have additional rra settings and when you imported the template it took these settings as well. If my maths servers me well, and you have 2 hourly 1 minute average rra, then that will be a lot of granular data collected, hence the large rrd file.

If you don't want such granular data then all you need to do is to un highlight the 1 minute averages in your data template. I think this will stop you graphing at these averages, I think. Have you got a test box you could test this on first. I don't want to give you advice that could cause problems for your current graphs.

If you import any future templates I would always go into the data template and check what rra settings its using before creating graphs. I have had the same issue as you, as in it created multiple graphs at different rra settings, but luckily it was in a test enviroment and I only wanted to check the templates worked before moving into production.

Cheers

Jay

[jay]

Be aware that you may have changed the step and heartbeat if you were using 1 minute polling.

See for more info http://forums.cacti.net/viewtopic.php?p=116403#116403

For 1 minute polling
step would be 60
heartbeat would be 120

for 5 minute polling
step would be 300
heartbeat would be 600

Not sure if this will cause problems or not.

Cheers

Jay

[albundy118]

hi @all,

I'm new in Cacti & Cisco IP SLA, but first of all, big thanks for that templates!
They make it really easy to gather IP SLA measurements and until now I got more or less everthing running.
But I still miss the OWD (one-way delay) measurements. Are they collected with snmp and I just don't find them in my graph?
I'm using IOS "12.4(2)T6".
sla monitor config

Code: Select all

ip sla monitor 1
 type jitter dest-ipaddr 10.0.0.2 dest-port 12345 codec g711ulaw
 tos 20
 threshold 160
ip sla monitor schedule 1 life forever start-time now

and show ip sla ... give some owd values

Code: Select all

Latency one-way time milliseconds
        Number of Latency one-way Samples: 11
        Source to Destination Latency one way Min/Avg/Max: 0/4/12 ms
        Destination to Source Latency one way Min/Avg/Max: 48/57/68 ms

Hope somebody could give a hint.

Thank in advanced
Florian

[pparcell]

Has anyone else noticed that when the RRD is created it is creating an RRD with all the DSs, including those that are not used? i.e. if you select IP HTTP, when it creates the RRD it is structuring it with DSs for all sources. Has anyone ran into this? It is a definite space killer since all the RRDs are the same size and have a lot of unused space.

My work around is breaking up the snmp template into a template for each type, and then separate graph and data templates and data queries. This resolves the issue, but I am not sure if there is a better way. I would have thought that if you selected only certain DSs in the data template of the data query it would create the RRD file with only those DSs. Any thoughts?

[sini]

Hi,

Please somebody tell me what does the minus values mean in SAA Jitter Statistics?

At the bottom of the graph there are
+Source > Dest (plus cur,avg,max values)
+Dest > Source (plus cur,avg,max values)
-Source > Dest (minus cur,avg,max values)
-Dest > Source (minus cur,avg,max values)

Thx

[sini]

Hi,

Please somebody tell me what does the minus values mean in SAA Jitter Statistics?

At the bottom of the graph there are
+Source > Dest (plus cur,avg,max values)
+Dest > Source (plus cur,avg,max values)
-Source > Dest (minus cur,avg,max values)
-Dest > Source (minus cur,avg,max values)

Thx

Ok, I know now

UDP Jitter Operation, Jitter Computation

If packets are sent with 10ms interval, positive
jitter means they have been received with more
than 10ms interval.
• Negative jitter means less than 10ms interval.
• Zero jitter means they are received with the same
inter-packet delay (the variance is zero).
• Jitter should remain as low as possible for real-
time traffic such as voice over IP.

[msab]

So far the IPSLA templates have been fantastic. One small issue. I added the mibs based on this

To get most out of the updated graph templates, I suggest importing the mibs for IP SLA. Attached is two mibs that needs to be copied to /mibs directory (On my Ubuntu linux, the location was /usr/share/snmp/mibs/). Once copied, add following lines in snmp.conf (might need to create the file if it does not exist, my snmp.conf resides in /usr/share/snmp/):

mibs +CISCO-SMI
mibs +CISCO-RTTMON-MIB

Once mibs and xml files are imported/copied go to 'New Graphs', Select your IP SLA configured router, Edit this host, Add Data Query, select 'Cisco - IP SLA Statistics', click on add. Now you can click on 'Create Graphs for this Host', and you should see a screen like below with informative description of the probe instead of just numbers. Also, the description will be carried on the actual graph tag entries.

I did exactly what was said but it did not seem to work. Are there any gotchas when doing this.

I had to create an snmp.conf file which I placed in the /usr/share/snmp dir.

Thanks
Mike

[sini]

Hi msab,

I use this template also with ubuntu with no problem.
Does the mibs loaded well?
Do you have any mib load errors in syslog?
Do an snmpwalk from cli and if any of the mibs not loaded well you should see some error in the output.

[chvp]

[quote="hwon"]To get most out of the updated graph templates, I suggest importing the mibs for IP SLA.

Hi
I'm trying to used the attached mibs but without any luck!
I've copied the files to the dir:
/usr/local/etc/snmp
I've checked that my machine can use these by doing a walk but in cacti my graph view is still the same! Do I need to re-install cacti or is there a better way to have cacti use these mibs!

I have no problem polling devices for data - using the Cisco IP SLA stuff!

[msab]

No error messages in /var/log/messages.

I am not sure where it looks for the snmp.conf file but thats what could be the problem.
My snmpwalk lives in /usr/bin/snmpwalk


snmpwalk seems to grab the correct enumerated types:
snmpwalk -v 1 -c anne 1.2.3.4 .1.3.6.1.4.1.9.9.42.1.2.1
CISCO-RTTMON-MIB::rttMonCtrlAdminOwner.100 = STRING:
CISCO-RTTMON-MIB::rttMonCtrlAdminTag.100 = STRING:
CISCO-RTTMON-MIB::rttMonCtrlAdminRttType.100 = INTEGER: echo(1)
CISCO-RTTMON-MIB::rttMonCtrlAdminThreshold.100 = INTEGER: 5000 milliseconds
CISCO-RTTMON-MIB::rttMonCtrlAdminFrequency.100 = INTEGER: 300 seconds
CISCO-RTTMON-MIB::rttMonCtrlAdminTimeout.100 = INTEGER: 5000 milliseconds
CISCO-RTTMON-MIB::rttMonCtrlAdminVerifyData.100 = INTEGER: false(2)
CISCO-RTTMON-MIB::rttMonCtrlAdminStatus.100 = INTEGER: active(1)
CISCO-RTTMON-MIB::rttMonCtrlAdminNvgen.100 = INTEGER: true(1)

I am using spine as my poller - could that be a problem.
Mike

[sini]

I also use spine so it is not problem.
And the Cisco IP SLA Statistics verbose data query in device menu also runs with no error?