juniper SSG host template

vpl · Post by **vpl** » Mon Nov 05, 2007 10:51 pm

I've taken the templates from the following posts and made them into one host template with a few modifications:

Memory, CPU, Sessions
http://forums.cacti.net/about3078.html
- changed memory to handle more than 300 (I think) MB. (My SSG520 was reporting NaN)

Tunnel Interfaces
http://forums.cacti.net/about11873.html
- added bits/sec with Total Bandwidth

Sub Interfaces
http://forums.cacti.net/about11064.html
- added bits/sec with Total Bandwidth

Place netscreen_interface.xml and netscreen_tunnel_interface.xml in <path_cacti>/resource/snmp_queries/ . Then just import the host template.

Any suggestions would be appreciated.

These are all working in version 0.8.7 with Spine 0.8.7.

*Update 13JAN08
- Using raw OIDs now.
- Added interface descriptions so interface names should show up now.
- Added another host template to the attachment where everything is renamed 'ScreenOS' instead of 'Netscreen' to reflect the fact that the devices are Juniper devices running ScreenOS, and not Netscreen anymore. Both templates work, it's a matter of personal preference. Pick one.

Much thanks goes to the creators of the original templates.

mcutting · Post by **mcutting** » Tue Nov 06, 2007 4:57 am

I seem to get dsname: cpu5, oid: SNMPv2-SMI::enterprises.3224.16.1.3.0, value: U appearing in the log. Have tried SNMP v1 and v2, but this doesn't seem to make any difference.

This MIB is also a standard one - any ideas why it is calling this ? This is the only MIB in my "mibs" directory - do I need the NetScreen one ?

Thanks

mcutting · Post by **mcutting** » Tue Nov 06, 2007 8:47 am

Ok - after a bit of fiddling around, I have found the following OID's, which are correct.

.1.3.6.1.4.1.3224.16.1.2.0 = Cpu Last 1 Minute
.1.3.6.1.4.1.3224.16.1.3.0 = Cpu Last 5 Minutes
.1.3.6.1.4.1.3224.16.1.4.0 = Cpu Last 15 Minutes
.1.3.6.1.4.1.3224.16.2.1.0 = Memory Allocated
.1.3.6.1.4.1.3224.16.2.2.0 = Memory Available
.1.3.6.1.4.1.3224.16.2.3.0 = Memory Fragmented
.1.3.6.1.4.1.3224.16.3.2.0 = Sessions Allocated
.1.3.6.1.4.1.3224.16.3.3.0 = Sessions Maximum
.1.3.6.1.4.1.3224.16.3.4.0 = Sessions Failed

You need to change the incorrect versions that are prefixed with "SNMPv2-SMI::enterprises.3224...". I have just done this with my templates, and now they all work.

AndySTT · Post by **AndySTT** » Thu Dec 27, 2007 10:01 pm

Great bundle of templates, I have been using these with a custom host template for some time.

The problem that I have is that the initial interogation of one of my NS208 clusters times out before the information has had a chance to be collected. I know the install and these templates work as I have them doing a fantastic job on approximately 100 other Firewalls.

The NS208 cluster in question is our main VPN tunnel concentrator for our main production site and has 80+ VPNs configured on it.

Is there any way known to increase the speed of the collection before the page times out? As you can image having the information on these tunnels is extremely important to us and our other option Solarwinds Orion won't perform proprietry mib indexing so is no use at all for this kind of work.

Many thanks for any help offered.

Andy

jay · Post by **jay** » Fri Dec 28, 2007 6:03 am

AndyStt

I had a similar problem with some chassis switches where it was timing out before it had scanned all the ports on the switch.

You need to do one or both of the following, I can't remember which one fixed the problem:-

Device - increase snmp timeout

Settings - Visual - Maximum JavaScript Rows - I set mine to 400 you may need more

Hope this helps

Jay

jay · Post by **jay** » Fri Dec 28, 2007 6:15 am

VPL

With regards to the sub interface naming. The graph template title should have the user per graph value ticked. Then within the data query the title should have the following |host_description| - NS Traffic - |query_nsIfName|. This should pick up the interface name in the graph title.

I will confirm this and get back to you.

Jay

AndySTT · Post by **AndySTT** » Mon Dec 31, 2007 5:32 am

jay wrote:Settings - Visual - Maximum JavaScript Rows - I set mine to 400 you may need more

Hope this helps

Jay

Can't find the Javascript rows field on the settings page, can anyone point me in the right direction?

jay · Post by **jay** » Mon Dec 31, 2007 5:39 am

Andy

See attached screenshot.

What version of Cacti are you using?? This setting may have changed in 8.7. I'm still using 8.6

Jay

AndySTT · Post by **AndySTT** » Mon Dec 31, 2007 6:29 am

Version 8.7b here.

jay · Post by **jay** » Mon Dec 31, 2007 6:37 am

I'm not sure how to change it with the new version of Cacti. What about under the graph creation tab. Is there anything which looks like it might allow you to show more rows when creating graphs??

Jay

AndySTT · Post by **AndySTT** » Mon Dec 31, 2007 6:44 am

jay wrote:I'm not sure how to change it with the new version of Cacti. What about under the graph creation tab. Is there anything which looks like it might allow you to show more rows when creating graphs??

Jay

Sorry Jay, been looking all over for 'Javascript'. Under 8.7b there is a 'Data Query Graph Rows -
The maximum number Data Query rows to place on a page per Data Query' which I've set to 1000, still no joy

The amount of time before I get the server error page doesn't appear to have changed by amending either this or the SNMP timeout value.

Still trying

jay · Post by **jay** » Mon Dec 31, 2007 6:53 am

Then I'm out of ideas

( I'm remember when I first added a large chassis switch into Cacti I had issues and changing one or both of these values helped. Then when I added another device which had even more slots/ports I was able to show all the items/rows.

This is the device with the most amount of ports I'm able to scan successfully.

SNMP - Interface Statistics (Verbose Query) Uptime Goes Backwards Success [2476 Items, 307 Rows]

I'm not aware of any limit and I'm sure there are people out there with bigger switches who are able to scan all the ports.

It might be worth creating a new post in the general help forum and see if someone with more expertise can help. Make sure you add the steps you have done - increase snmp timeout and changing the graph rows.

Jay

AndySTT · Post by **AndySTT** » Mon Dec 31, 2007 11:38 am

Thanks for trying Jay, will do

vpl · Post by **vpl** » Sun Jan 13, 2008 2:22 pm

Thanks jay. Your suggestion regarding the naming worked. I added |query_nsIfName| to all the regular interface graph templates and |query_ifName| to all the tunnel interface graph templates. Interface and VPN tunnel names are showing up great now.

I've updated the attached host template to include the raw OIDs and the above naming fixes. And it looks like AndySTT got his question answered http://forums.cacti.net/about25052.html.

Eli Klein · Post by **Eli Klein** » Tue Jul 22, 2008 10:41 am

Nice work on this, thanks!

Cacti

juniper SSG host template

juniper SSG host template

Tunnel info timeout

Who is online