Nov 5, 2011 1:52 PM
device report errors
-
Like (0)
What version of Zenoss on what platform? I've seen you post a bunch, and it seems like there's a lot sort of broken with your install, but I don't recall all the details for each thread. It's always a good idea to put full info in a thread so it's self contained.
--
James Pulver
ZCA Member
LEPP Computer Group
Cornell University
Sorry, This is Zenoss 3.2.1 on RHEL5 64 bit
I figured this out. It was due to a broken ZenPack
Which one? Knowing which is broken, could save others your same pain in the future, or alert the original author that something is up.
Its the Community F5 zenpack. Turns out I broke the ZenPack, or maybe its just not completely compatible with this version (3.2.1) Once I set the device class to /Network/f5 it would not model and I would see things like
AttributeError: 'BigipLtm' object has no attribute 'primaryAq'
in zenhub. I've since re-installed this zenpack and to get around the "not modeling problem" I used zendmd to to this:
>>> d.setPerformanceMonitor('localhost')
>>> commit()
>>> d.collectDevice
now the Big IP devices will Model and my device reports work.
Well there is some Irony.... Thats my ZenPack
I'm just now (as of this week) testing it on 3.2.1, previously I only tested it on 3.0.3 and 3.1. If there is a bug, I'm very much interestedin resolving it.
When you say "Once I set the device class to /Network/f5" what exactly does that mean? Did you move an existing device into that device class, or something else?
I don't recall having any references to primaryAq in that code, but I'll search through it again tonight.
And steps to resolve were:
1) Re-Install The Pack
2) Run the 3 line DMD script above?
I have another f5 to add. I'll do it right now and pay closer attention and post a again in a bit.
yes, all I did to fix it was
d = find('mydevice')
d.setPerformanceMonitor('localhost')
commit()
d.collectDevice
Again, there's a good chance I screwed this up somehow. I'll do it again now and see if I can't break it again
This time I added the F5 by hostname and only set the SNMP community name.
It initially had no device class but it was reporting basic info like interfaces
Next I changed the device class to /Network/f5 - get the yellow warning banner on the page that says "failed to move devices". Returning to the infrastructure page I see that it is /Network/f5 (despite the error message)
Next I bind both of the BigIP templates
Now try to model the device, you get
Traceback (most recent call last): File "/opt/zenoss/Products/ZenUI3/browser/streaming.py", line 44, in __call__ self.stream() File "/opt/zenoss/Products/ZenUI3/browser/command.py", line 144, in stream device.collectDevice(REQUEST=self.request, write=self.write) File "/opt/zenoss/Products/ZenModel/Device.py", line 1824, in collectDevice perfConf.collectDevice(self, setlog, REQUEST, generateEvents, AttributeError: 'NoneType' object has no attribute 'collectDevice'
This time I set the collector via the UI (so UI and zendmd work here) and then modeled the device and it worked.
After a successfull model, I still have no data. I notice the SNMP community name has reverted to the default.
Set the community name and remodel
Viola!
hope that helps
Well the good news (for me), is I don't think this is direclty related to the ZenPack itself. The traceback all references core modules and no Zenpack modules.
All that aside, I'm a bit baffled by your approach of adding new devices (or at least this device).
I'm assuming you are adding the device via the Web UI? If so won't don't you simply choose the proper device class on the device add screen rather than attempting to move it there after the fact?
If you had the device directly, or drag/drop into the /Network/f5 class, the appropriate templates should automatically get inherited from the device class, avoiding the need to manually bind them.
All and All I think your install is clean, and you just need to refine your process for adding new devices.
No doubt, I am still learning Zenoss, thanks for your help and patience.
Even though this appears to be configured correctly I'm getting messages like:
Error reading value for "TotalConnections" on x.mydevice.com (oid .1.3.6.1.4.1.3375.2.2.10.2.3.1.11 is bad)
yet I can snmpwalk that OID just fine.
Try the following:
1) Log into your Zenoss server via SSH
2) change to the Zenoss User
3) Run the following (going from memory, not in front of zeonss box):
zenperfsnmp run -d x.mydevice.com - v10
This should produce some really verbose output and ideally give a little more insight on whats up.
hrmm, no smoking gun. This doesnt seem right though
2011-11-07 16:40:03,086 INFO zen.zenperfsnmp: success:0 fail:0 pending:0 todo:0
Yeah seems off... Where did you see "Error reading value for "TotalConnections" on x.mydevice.com (oid .1.3.6.1.4.1.3375.2.2.10.2.3.1.11 is bad)" message?
Have you restarted Zenoss since re-installing the zenpack?
Its in the event log but I notice the event is a few hours old and it did not repeat. Now its just doing nothing
Follow Us On Twitter »
|
Latest from the Zenoss Blog » | Community | Products | Services Resources | Customers Partners | About Us | ||
Copyright © 2005-2011 Zenoss, Inc.
|
||||||||