Switch information Nortel Networks Layer 2-3 GBE Switch Module(Copper)

[Duarte Franco]

Hi

I am trying to get the switch information of an IBM Bladecenter. The switch
modules are Nortel Networks Layer 2-3 GBE Switch Module(Copper). OID is
1.3.6.1.4.1.1872.1.18.1

Is there any way to make the data acquisition of this type of switch? In the "IP
Device Info" i saw 0 ports an the cricket have no data.

There is some workaround for this type of switch?

Best Regards

[Morten Brekkevold]

Post by Duarte Franco
Hi
I am trying to get the switch information of an IBM Bladecenter. The switch
modules are Nortel Networks Layer 2-3 GBE Switch Module(Copper). OID is
1.3.6.1.4.1.1872.1.18.1
Is there any way to make the data acquisition of this type of switch? In the "IP
Device Info" i saw 0 ports an the cricket have no data.
There is some workaround for this type of switch?

We've never tested with this switch, so I couldn't say off the top of my
head.

Which version of NAV are you running? Are other switches being detected
correctly by NAV? Are there any error messages related to this switch in
ipdevpoll.log?

[Duarte Franco]

Hi Mortan

Im running version 3.8.6

In the ipdevpoll.log i only have this error:

[ERROR jobs.jobhandler] [inventory se03bc05sw02.priv.fccn.pt] Job 'inventory'
for se03bc05sw02.priv.fccn.pt aborted: SNMP request timed out

But the switch respond to snmp querys perfectly from the console:

snmpwalk -v2c -c community SWITCH
SNMPv2-MIB::sysDescr.0 = STRING: BNT Layer 2/3 Copper Gigabit Ethernet Switch
Module for IBM BladeCenter
SNMPv2-MIB::sysObjectID.0 = OID: SNMPv2-SMI::enterprises.1872.1.18.1
DISMAN-EVENT-MIB::sysUpTimeInstance = Timeticks: (145678900) 16 days, 20:39:49.00

[Morten Brekkevold]

Post by Duarte Franco
Hi Mortan
Im running version 3.8.6
[ERROR jobs.jobhandler] [inventory se03bc05sw02.priv.fccn.pt] Job 'inventory'
for se03bc05sw02.priv.fccn.pt aborted: SNMP request timed out

There a are a few known issues with SNMP timeouts when using ipdevpoll:

1. Your device may be slow at generating responses for GetBulk requests.
This has at least been demonstrated to be problem with some devices
from other vendors.

2. You may want to patch your TwistedSNMP library as described in NAV's
NOTES file. We submitted a patch to TwistedSNMP to fix a problem
with the timeout not increasing as it should between retries, but
we've never received a response. The patch can be found in NAV's
contrib directory.


If you've applied the patch and are still getting timeouts, you can set
up some more logging in ipdevpoll so we can at least see which plugin
times out. Create a file in NAV's config directory called
`logging.conf`, and add the following three lines:

[levels]
nav.ipdevpoll = INFO
nav.ipdevpoll.jobs.jobhandler = DEBUG

[Duarte franco]

Hi Mortan

I have made the patch to TwistedSNMP library but i still receiving the
timeout.

Increasing the verbosity of the logs i receive:


2011-07-27 12:56:31,538 [DEBUG jobs.jobhandler] [profiling SWITCH] Job
'profiling' initialized with plugins: ['oidprofiler']
2011-07-27 12:56:31,538 [DEBUG jobs.jobhandler] [profiling SWITCH]
AgentProxy
created for SWITCH:
AgentProxy(SWITCH_IP,161,community,v2c,<twistedsnmp.snmpprotocol.SNMPProtoco
l
instance at 0xa9c760c>)
2011-07-27 12:56:31,539 [DEBUG jobs.jobhandler] [profiling SWITCH] Plugin
oidprofiler wouldn't handle SWITCH
2011-07-27 12:56:31,539 [DEBUG jobs.jobhandler] [profiling SWITCH] No
plugins
for this job
2011-07-27 12:56:31,539 [INFO schedule.netboxjobscheduler] [profiling
SWITCH]
Next 'profiling' job for SWITCH will be in 299 seconds (2011-07-27
13:01:31.539127)

Apparently the oidprofiler not recognize the router OID.

-----Original Message-----
From: Morten Brekkevold [mailto:morten.brekkevold-***@public.gmane.org]
Sent: quarta-feira, 27 de Julho de 2011 08:08
To: Duarte Franco
Cc: nav-users-***@public.gmane.org
Subject: Re: Switch information Nortel Networks Layer 2-3 GBE Switch
Module(Copper)

On Tue, 26 Jul 2011 11:15:18 +0000 (UTC) Duarte Franco

Post by Duarte Franco
Hi Mortan
Im running version 3.8.6
[ERROR jobs.jobhandler] [inventory se03bc05sw02.priv.fccn.pt] Job 'inventory'
for se03bc05sw02.priv.fccn.pt aborted: SNMP request timed out

There a are a few known issues with SNMP timeouts when using ipdevpoll:

1. Your device may be slow at generating responses for GetBulk requests.
This has at least been demonstrated to be problem with some devices
from other vendors.

2. You may want to patch your TwistedSNMP library as described in NAV's
NOTES file. We submitted a patch to TwistedSNMP to fix a problem
with the timeout not increasing as it should between retries, but
we've never received a response. The patch can be found in NAV's
contrib directory.


If you've applied the patch and are still getting timeouts, you can set up
some more logging in ipdevpoll so we can at least see which plugin times
out. Create a file in NAV's config directory called `logging.conf`, and add
the following three lines:

[levels]
nav.ipdevpoll = INFO
nav.ipdevpoll.jobs.jobhandler = DEBUG


--
Morten Brekkevold
UNINETT

[Morten Brekkevold]

Post by Duarte franco
I have made the patch to TwistedSNMP library but i still receiving the
timeout.

[snip]

Post by Duarte franco
2011-07-27 12:56:31,539 [DEBUG jobs.jobhandler] [profiling SWITCH] Plugin
oidprofiler wouldn't handle SWITCH

[snip]

Post by Duarte franco
Apparently the oidprofiler not recognize the router OID.

No, the oidprofiler's job is to test a device for compatibility with
various statistics-related SNMP OIDs that Cricket will use. It rejected
the router on this run because its OID profile was already up-to-date.

I'm more interested in the log messages from the 'inventory' job for
your router, which was the job that was timing out in your original log
excerpt.

[Duarte franco]

Hi Morton, pleased see you again.

Here is my inventory log for this switch. Thanks for your help.


2011-08-12 11:03:28,267 [DEBUG jobs.jobhandler] [inventory SWITCH] Job
'inventory' initialized with plugins: ['dnsname', 'typeoid', 'modules',
'interfaces', 'bridge', 'dot1q', 'ciscovlan', 'prefix', 'lastupdated']
2011-08-12 11:03:28,267 [DEBUG jobs.jobhandler] [inventory SWITCH]
AgentProxy created for SWITCH:
AgentProxy(IP,161,COMUNITY,v2c,<twistedsnmp.snmpprotocol.SNMPProtocol
instance at 0xa9c282c>)
2011-08-12 11:03:28,268 [DEBUG jobs.jobhandler] [inventory SWITCH] Plugins
to call:
DnsName,TypeOid,Modules,Interfaces,Bridge,Dot1q,CiscoVlan,Prefix,LastUpdated
2011-08-12 11:03:28,268 [INFO jobs.jobhandler] [inventory SWITCH] Starting
job 'inventory' for SWITCH
2011-08-12 11:03:28,268 [DEBUG jobs.jobhandler] [inventory SWITCH] Now
calling plugin: nav.ipdevpoll.plugins.dnsname.DnsName(u'SWITCH')
2011-08-12 11:03:28,271 [DEBUG jobs.jobhandler] [inventory SWITCH] Now
calling plugin: nav.ipdevpoll.plugins.typeoid.TypeOid(u'SWITCH')
2011-08-12 11:03:28,318 [DEBUG jobs.jobhandler] [inventory SWITCH] Now
calling plugin: nav.ipdevpoll.plugins.modules.Modules(u'SWITCH')
2011-08-12 11:03:28,414 [INFO plugins.modules.modules] [inventory SWITCH] No
chassis found
2011-08-12 11:03:28,415 [DEBUG jobs.jobhandler] [inventory SWITCH] Now
calling plugin: nav.ipdevpoll.plugins.interfaces.Interfaces(u'SWITCH')
2011-08-12 11:03:29,021 [DEBUG jobs.jobhandler] [inventory SWITCH] Now
calling plugin: nav.ipdevpoll.plugins.bridge.Bridge(u'SWITCH')
2011-08-12 11:03:35,684 [DEBUG jobs.jobhandler] [inventory SWITCH] Plugin
nav.ipdevpoll.plugins.bridge.Bridge(u'SWITCH') reported a timeout
2011-08-12 11:03:35,685 [DEBUG jobs.jobhandler.timings] [inventory SWITCH]
Job 'inventory' timings for SWITCH:
2011-08-12 11:03:35,686 [ERROR jobs.jobhandler] [inventory SWITCH] Job
'inventory' for SWITCH aborted: SNMP request timed out
2011-08-12 11:03:35,686 [INFO schedule.netboxjobscheduler] [inventory
SWITCH] Next 'inventory' job for SWITCH will be in 386 seconds (2011-08-12
11:10:01.686431)




-----Original Message-----
From: Morten Brekkevold [mailto:morten.brekkevold-***@public.gmane.org]
Sent: terça-feira, 9 de Agosto de 2011 12:01
To: Duarte franco
Cc: nav-users-***@public.gmane.org
Subject: Re: Switch information Nortel Networks Layer 2-3 GBE Switch
Module(Copper)

Post by Duarte franco
I have made the patch to TwistedSNMP library but i still receiving the
timeout.

[snip]

Post by Duarte franco
2011-07-27 12:56:31,539 [DEBUG jobs.jobhandler] [profiling SWITCH]
Plugin oidprofiler wouldn't handle SWITCH

[snip]

Post by Duarte franco
Apparently the oidprofiler not recognize the router OID.

No, the oidprofiler's job is to test a device for compatibility with various
statistics-related SNMP OIDs that Cricket will use. It rejected the router
on this run because its OID profile was already up-to-date.

I'm more interested in the log messages from the 'inventory' job for your
router, which was the job that was timing out in your original log excerpt.

--
Morten Brekkevold
UNINETT

p.s. Sorry about the late reply, I've been on vacation.

[Morten Brekkevold]

Post by Duarte franco
Here is my inventory log for this switch. Thanks for your help.
2011-08-12 11:03:35,684 [DEBUG jobs.jobhandler] [inventory SWITCH] Plugin
nav.ipdevpoll.plugins.bridge.Bridge(u'SWITCH') reported a timeout

Hi Duarte,

There's the line I was looking for. The Bridge plugin is to blame. I
think I've seen this happen before, and although I have my suspicions
about what's wrong, I'm not entirely sure.

If you would also add debug logging to that plugin, that would be very
helpful. Add the following line to logging.conf:

nav.ipdevpoll.plugins.bridge = DEBUG