be more flexable with oid/regex output in XML files

sodium · Joined: 01 Feb 2006 Posts: 24

I have seen several MIB's where an index is used to point to other locations in the mib with a little extension (of 2 numbers)

example: (juniper mib)

.1.3.6.1.4.1.2636.5.1.1.2.1.1.1.14.<index>.<peer_index_#>

and now you have to use <index> on the info you want but now with an addition of 2 numbers

like:

.1.3.6.1.4.1.2636.5.1.1.2.6.2.1.7.<peer_index_#>.1.1
.1.3.6.1.4.1.2636.5.1.1.2.6.2.1.8.<peer_index_#>.1.1
.1.3.6.1.4.1.2636.5.1.1.2.6.2.1.9.<peer_index_#>.1.1
.1.3.6.1.4.1.2636.5.1.1.2.6.2.1.10.<peer_index_#>.1.1

if we could use back references from regex in XML like this:

<interface>
<name>Juniper - BGP Peer Prefix Stats</name>
<oid_index>.1.3.6.1.4.1.2636.5.1.1.2.1.1.1.14</oid_index>
<oid_index_parse>OID/REGEXP:.*\.([0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3})\.1\.([0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3})\.([0-9]{1,2})$</oid_index_parse>
<index_order>Index</index_order>
<index_order_type>numeric</index_order_type>
<index_title_format>|chosen_order_field|</index_title_format>

<fields>
<Index>
<name>Index</name>
<method>walk</method>
<source>value</source>
<direction>input</direction>
<oid>.1.3.6.1.4.1.2636.5.1.1.2.6.2.1</oid>
</Index>

<jnxBgpM2PrefixInPrefixes>
<name>Prefixes In</name>
<method>walk</method>
<source>value</source>
<direction>output</direction>
<oid>.1.3.6.1.4.1.2636.5.1.1.2.6.2.1.7.$3.1.1</oid>
</jnxBgpM2PrefixInPrefixes>
<jnxBgpM2PrefixInPrefixesAccepted>
<name>Accepted Prefixes</name>
<method>walk</method>
<source>value</source>
<direction>output</direction>
<oid>.1.3.6.1.4.1.2636.5.1.1.2.6.2.1.8.$3.1.1</oid>
</jnxBgpM2PrefixInPrefixesAccepted>
<jnxBgpM2PrefixOutPrefixes>
<name>Outbound Prefixes</name>
<method>walk</method>
<source>value</source>
<direction>output</direction>
<oid>.1.3.6.1.4.1.2636.5.1.1.2.6.2.1.10.$3.1.1</oid>
</jnxBgpM2PrefixOutPrefixes>
</fields>
</interface>

I do not have to do this in a seperate perl script...

and the Centillion MIB has the same problem -- so I think it's more common....

twelzy · Posted: Wed Jun 21, 2006 4:09 pm Post subject:

Hi sodium!

That's one of the features that I requested in 'Improve the "power" of REGEXPs in Data Queries' (http://forums.cacti.net/viewtopic.php?t=13949)!
(No answer until now...)

Still waiting...

(And sorry for my poor English...)

rony · Posted: Thu Jun 22, 2006 9:27 am Post subject:

I don't know how we can make it "better", as we are using the regex functions that PHP provides.

sodium · Joined: 01 Feb 2006 Posts: 24

danbeck · Joined: 20 Jun 2006 Posts: 30 Location: Huntsville, Alabama

rony · Posted: Thu Jun 22, 2006 5:18 pm Post subject:

twelzy · Posted: Thu Jun 22, 2006 7:10 pm Post subject:

Hi sodium, rony and danbeck,

The actual regexp functions used in <CACTI_PATH>/lib/data_query.php file are POSIX Regular Expressions.

I think that both POSIX Regular Expressions functions and PCRE (Perl-Compatible Regular Expression) functions could resolve this issue.

The problem here is that with the actual <CACTI_PATH>/lib/data_query.php file we can't do:

In a previous post of mine (http://forums.cacti.net/viewtopic.php?t=13949) I give this example:

twelzy · Posted: Thu Jun 29, 2006 3:04 pm Post subject:

Any feedback...?!

Thanks in advance!
(And sorry for my poor English...)

sodium · Joined: 01 Feb 2006 Posts: 24

The silence is deafining...

TheWitness · Posted: Tue Jul 11, 2006 9:28 pm Post subject:

It's just that the developers are indisposed at the moment. I can't seem to get any time for Cacti lately. I hope that changes soon....

TheWitness

sebastianw · Joined: 31 Jul 2007 Posts: 5

Hi! I also asked for a feature like that in

http://forums.cacti.net/viewtopic.php?t=22586

I use Cacti with Cisco routers and I would greatly benefit from a feature which would let me use a suffix for some OIDs (see my original post for an example).

sh0x · Joined: 30 Aug 2007 Posts: 14

I could use this feature as well. I'm finding this to be a common issue with many Cisco OIDs and would imagine its the same for other vendors MIBs as well.

For example I'm trying to poll frame-relay statistics and my indexes are:

14.100
14.200
19.100
19.200

Then I poll a bunch of OIDs and one of my OIDs adds a .1 to the end so I can't poll that particular OID.

14.100.1
14.200.1
19.100.1
19.200.1

I'm finding this more of an issue when I try to reference another OID for a descriptive information. I'm new to regex though. Can I do a regex to ignore the .1 if it exists, but to still work if it doesn't have a .1?

Thanks,
sh0x

Syngress · Joined: 24 Apr 2007 Posts: 13

Hi,

I am having the same issue as the original poster in that the OID I want use has a suffix to the index

http://forums.cacti.net/viewtopic.php?t=23596 is where I posted my original issue, any idea if this will get fixed as a lot of our new network equipment (Juniper & Foundry) use these sort of OID's.

Syngess

Ragnar · Joined: 04 Aug 2005 Posts: 5

also requesting this feature, we have quite a few pieces of hardware (Cisco routers, MetaSwitch softswitch) where the developers have suffixed an OID, instead of prefixing into a table like they should have...

Anyone have a workaround/patch for this? We only have a single digit to deal with on the end (.1 = current 5 mins, .2 = prev 5 mins, .3 = current 15 mins, etc), my cacti-fu is weak though, unless it's hacking together composite bandwidth graphs...

Sample of snmpwalk output of the OID's I'm trying to graph:

TheWitness · Posted: Thu Nov 15, 2007 4:15 pm Post subject:

Does anyone have a patch? It won't make it into Saturday's release, but...

TheWitness