Help to monitor an McData SAN Direcor ED-10000M

[anovelli]

Hello!

This is my very first post here, so, s0rry if it is not the right place to put this message...

I have dozen of SAN Directors to monitor with cacti and most of them is the McData ED-10000M.
I'm not an expert in SNMP and cacti, but i'm good at php programming.

Tried some graph templates found here but no one seems to work.

Can you, cacti gurus place me a direction ? from where to start and understand how to implement this?
thank you!

[gandalf]

Please search the Script and Templates forum
R.

[anovelli]

Hello gandalf,

I've made the search you suggested me and found a good post about it:

http://forums.cacti.net/about5747.html&highlight=mcdata

So i imported and copied the files to the right place, but i can't graph port traffic, in other hand, if i choose another type of data like "Class 1 Frame", i can graph it, but nothing about traffic.

as i Read in another post (http://forums.cacti.net/viewtopic.php?t ... ght=mcdata) there are some differences between mcdata models in the OID number, so, where can i modify it in the template in order to correct this issue?

Thank you!

[gandalf]

To understand templating, please read 1st link of my sig.
R.

[anovelli]

Hello Gandalf,

Thank you for your help.

After some research on SNMP basics, i found the data that i want to graph.

here is the Textual OID that i found with snmpwalk
EOSN-MIB::bbPerfPortTxRate
This OID means the numeric oid .1.3.6.1.4.1.289.2.2.3.1.6.9.2.1.1.5

when i run

Code: Select all

shell> snmpwalk -c public -v 2c -m ALL <host> .1.3.6.1.4.1.289.2.2.3.1.6.9.2.1.1.5.1

i receive the information about the port #1
(textual) EOSN-MIB::bbPerfPortTxRate.1 = Gauge32: 133812 Bps
(numeric [-On]) .1.3.6.1.4.1.289.2.2.3.1.6.9.2.1.1.5.1 = Gauge32: 354388 Bps
the value changed because it is the expected
so, the value brings to me the "Bps" suffix.
in the cacti verbose query:

Code: Select all

+ Executing SNMP walk for data @ '.1.3.6.1.4.1.289.2.2.3.1.6.9.2.1.1.6'
+ Found item [hrPortrx='860909'] index: 1 [from value]

Note that the value is much different in size, so i think that the "Bps" string suffix is the offender (maybe there is a data conversion routine in cacti wich does not strip that suffix ? )

Could you show me how to strip that, maybe inside the template .xml?
thank you!

[gandalf]

You used different OIDs in your example. Was this on purpose?
R.

[anovelli]

One is for Tx and other is for Rx, sorry for the mess, but the data is the same.

I found another oid wich brings to me the same data, but Hex coded, wich one is easier to modify in cacti ?

[anovelli]

a little bit of improvement....

Following the debug tutorial (also in your signature, gandalf ), i got somethig curious.

Look:

Code: Select all

 /usr/local/spine/spine --verbosity=5 19 19

Result transcription:
...
02/18/2011 09:22:17 PM - SPINE: Poller[0] Host[19] TH[1] DS[107] SCRIPT: /usr/local/bin/php -q /usr/local/apache2/htdocs/cacti/scripts/query_mds9020_san_portinfo.php CNX10 public 2 get rxelement 84, output: 3.7710229732296E+14
02/18/2011 09:22:17 PM - SPINE: Poller[0] Host[19] DEBUG: The NIFTY POPEN returned the following File Descriptor 6
02/18/2011 09:22:18 PM - SPINE: Poller[0] Host[19] TH[1] DS[107] SCRIPT: /usr/local/bin/php -q /usr/local/apache2/htdocs/cacti/scripts/query_mds9020_san_portinfo.php CNX10 public 2 get txelement 84, output: 1.6289557235313E+15
...

Can someone tell me how/if cacti deals with that float number (above, in bold)?
After a lot of reading, I think that I'm closer to solve my issue.

Thanks in advance!

[anovelli]

After reseting the counter on the port (in the switch admin), I was able to graph correctly,

Someone know how can i resolve this ? I'm sure that when the counter become too big again cacti will stop to graph.

thanks!