Re: CSS11000

I have a half-working Cisco CSM module for genDevConfig - the CSM indexes the same way - ie with strings encoded as ASCII OIDs.

The CSS module (which was written long ago) writes out the OIDs in full which is not particularly readable nor clever. For the CSM module (which I started much later) I tried a different route - I added a perl subroutine ascii2oid("string") to Cricket so that it was possible to use config like...

target ...
    ...
    module = 4
    inst = %module%.ascii2oid("ws2-public").80

with cricket being such a hack, syntactically it wasn't quite that straight forward but it was certainly much easier to read the configurations. The associated CSM module then wrote out configs using the ascii2oid.

I can dig up the subroutine and some examples if anyone is interested but I don't have a CSM to develop against anymore. 

Cheers


----- Original Message ----
From: Brent Van Dussen <[email protected]>
To: Rich West <[email protected]>; [email protected]
Sent: Saturday, 27 January, 2007 6:45:19 PM
Subject: Re: [cricket-users] CSS11000


I recently put a cricket installation in place for a client that has 
CSS11000's.  I was partially successful at using a map() to make it easier 
to create targets.

It's funny how different each SLB vendor implements their ifindex-like 
mapping in SNMP.  For the CSS, the content/service rule names are converted 
to their ASCII decimal equivalents one letter a time to formulate the 
OID.  A foundry will use the IP address and port of a VIP/RIP and 
incorporate that into the OID, alteons are just simple and have a number 
assigned to each group/virt/real which is also the snmp index number, etc...

There are OID's that map the content rule names and the service rule names 
to their ACSCII equivalent's so those are easy to map out but what got me 
was that you can't use multiple maps to make up an snmp string.

# CSS 11000 OID's
oid     apCntName       .1.3.6.1.4.1.9.9.368.1.16.4.1.2
oid     apCntHits       .1.3.6.1.4.1.9.9.368.1.16.4.1.18
oid     apCntByteCount  .1.3.6.1.4.1.9.9.368.1.16.4.1.25
oid     apCntFrameCount .1.3.6.1.4.1.9.9.368.1.16.4.1.26
oid     apSvcNames      .1.3.6.1.4.1.9.9.368.1.15.2.1.1
oid     apCntsvcHits    .1.3.6.1.4.1.9.9.368.1.18.2.1.4

datasource cssservicehits
   ds-source     = snmp://%snmp%/apCntsvcHits.%inst2%.%inst%


targettype service-hits
         ds      = "cssservicehits"
         view    = "hits: cssservicehits"


target --default--
   target-type = service-hits
   inst          = map(service-name)
   inst2         = map(content-name)

map service-name
   base-oid      = apSvcNames
   match         = %service-name%

map content-name
   base-oid      = apCntName
   match         = %content-name%

I tried the above config but found out that the perl code that does the 
mapping will only look for "inst" to trigger the mapping so any other 
variable name will not work.  I think a cool upgrade (based on funky OID 
strings like the CSS uses) would be to allow inst[0-9].

So I ended up just commenting out the inst2 mapping section and just adding 
it to each individual target config instead so I didn't have to worry about 
changing the ds-source line.

target lxksease1e_downloads_http
   service-name = lxksease1e_downloads_http
   inst2 = 
8.76.52.95.79.119.110.101.114.14.100.111.119.110.108.111.97.100.115.95.104.116.116.112

Naming your rules really short in foresight will make your snmp monitoring 
job a little bit easier :)

Hit me up if you have any questions about the above...Also I'm not 
extrodinarily talented at PERL hacking...can someone see if making multiple 
mapping instances work is easily doable?

-Brent



At 11:33 AM 12/6/2006, Rich West wrote:
>I came across a couple of postings in the archives related to people who
>have managed to get cricket set up to monitor their CSS11000's, but the
>information was somewhat limited.
>
>Does anyone have any working examples of a configuration for a
>CSS11000?  Also, if you do, can you explain how you managed to determine
>what entries were related to what custom rules?  The problem with the
>CSS11000 SNMP output is that each system is custom.. it's depends upon
>the local configuration and local rules for the particular CSS.
>
>However, if I could just nail down the mapping of the output from SNMP
>from a CSS and the respective rules, I think I might be able to make
>some headway. :)
>
>-Rich
>
>-------------------------------------------------------------------------
>Take Surveys. Earn Cash. Influence the Future of IT
>Join SourceForge.net's Techsay panel and you'll get the chance to share your
>opinions on IT & business topics through brief surveys - and earn cash
>http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>_______________________________________________
>cricket-users mailing list
>[email protected]
>https://lists.sourceforge.net/lists/listinfo/cricket-users


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
cricket-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/cricket-users
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV