TECHZEN Zenoss User Community ARCHIVE  

SNMP Collector Creation Failed

Subject: SNMP Collector Creation Failed
Author: Kai H
Posted: 2018-04-04 14:36


Hello All,

I am running into the following SNMP issue when modeling a device (linux server) from a remote collector. We are currently running Zenoss 6.1.1 (Service Dynamics) in a test environment. When modeling the device, the following error is received:

2018-04-04 18:12:47,640 INFO zen.ZenModeler: Connecting to localhost:8789
2018-04-04 18:12:47,705 INFO zen.ZenModeler: Connected to the zenhub/0 instance
2018-04-04 18:12:47,706 INFO zen.ZenModeler: Getting configuration from ZenHub...
2018-04-04 18:12:48,080 INFO zen.ZenModeler: Starting modeling in 0 seconds.
2018-04-04 18:12:48,080 INFO zen.ZenModeler: Starting collector loop #001...
2018-04-04 18:12:48,080 INFO zen.ZenModeler: Collecting for device 
2018-04-04 18:12:48,080 INFO zen.ZenModeler: Got 1 devices to be scanned during collector loop #001
2018-04-04 18:12:48,220 INFO zen.ZenModeler: Filled collection slots for 1 of 1 devices during collector loop #001
2018-04-04 18:12:48,221 INFO zen.ZenModeler: skipping WMI-based collection, PySamba zenpack not installed
2018-04-04 18:12:48,221 INFO zen.ZenModeler: Collect on device  for collector loop #001
2018-04-04 18:12:48,561 INFO zen.ZenModeler: No Python plugins found for
2018-04-04 18:12:48,565 INFO zen.ZenModeler: No command plugins found for 
2018-04-04 18:12:48,570 INFO zen.ZenModeler: Skipped collection of 
2018-04-04 18:12:48,570 WARNING zen.ZenModeler: SNMP collector creation failed
2018-04-04 18:12:48,575 INFO zen.ZenModeler: No portscan plugins found for 
2018-04-04 18:12:48,575 INFO zen.ZenModeler: Scan time: 0.49 seconds for collector loop #001
2018-04-04 18:12:48,575 INFO zen.ZenModeler: Scanned 1 of 1 devices during collector loop #001
2018-04-04 18:12:48,576 INFO zen.ZenModeler: Daemon ZenModeler shutting down

To troubleshoot, the delegate host (collector) has been rebuilt several times, restarted all collector daemons, and moved the device to the localhost to test modeling. When the device is on the localhost collector, the device is modeled without issue. When moved back to the remote collector, the error listed above is received. 

Also, all ports and firewalls have been checked. I am able to connect to all ports to the Resource manager and vice versa with no issues. I have also confirmed that all traffic is flowing between the collector, the monitored devices, and back to the Resource Manager. Any help would be greatly appreciated! Thank you!


Subject: RE: SNMP Collector Creation Failed
Author: Jane Curry
Posted: 2018-04-05 03:57

Can you run zenmodeler stand-alone against your test device with full logging?  You will need to attach to the zenmodeler container and then run:

zenmodeler run -v 10 -d <device name>  > /tmp/zm.out 2>&1

and then inspect /tmp/zm.out.  You might find more clues there.

Cheers,
Jane

------------------------------
Jane Curry
Skills 1st United Kingdom
jane.curry@skills-1st.co.uk
------------------------------


Subject: RE: SNMP Collector Creation Failed
Author: Kai H
Posted: 2018-04-05 13:30

I went ahead and attached to the "zenmodeler" and ran the listed zenmodeler command. I attached the output to this post for reference because it is quite long. If it is easier to post the whole log here rather than attach it please let me know. There was an interesting line at the bottom of the log:

DEBUG zen.pbclientfactory: clientConnectionLost [Failure instance: Traceback (failure with no frames): <class 'twisted.internet.error.ConnectionLost'>: Connection to the other side was lost in a non-clean fashion: Connection lost.
]

Not sure why the connection would be lost or timeout as there should be nothing that is stopping that traffic. So I am at a bit of a loss here. Any help would be greatly appreciated.

------------------------------
Kai H
------------------------------

Attachments:

zenmodeler_collector_troubleshoot.txt



Subject: RE: SNMP Collector Creation Failed
Author: Jane Curry
Posted: 2018-04-06 02:57

Is the zenperfsnmp daemon actually running on that collector?  I'm guessing not.

Attach to the zope container and do a "zenoss status".

Attach to the zenperfsnmp container and inspect /opt/zenoss/log/zenperfsnmp.log.

Cheers,
Jane

------------------------------
Jane Curry
Skills 1st United Kingdom
jane.curry@skills-1st.co.uk
------------------------------


Subject: RE: SNMP Collector Creation Failed
Author: Kai H
Posted: 2018-04-06 11:25

So I attached to the "zenperfsnmp" container and found that the daemon is not started. See below:

Daemon: zeneventserver not running
Daemon: zopectl program running; pid=174
Daemon: zenhub not running
Daemon: zenjobs not running
Daemon: zeneventd not running
Daemon: zenping not running
Daemon: zensyslog not running
Daemon: zenstatus not running
Daemon: zenactiond not running
Daemon: zentrap not running
Daemon: zenmodeler not running
Daemon: zenperfsnmp not running
Daemon: zencommand not running
Daemon: zenprocess not running
Daemon: zenpython not running
Daemon: zenjmx not running
Daemon: zenjserver not running
Daemon: zenwebtx not running
Daemon: zenpropertymonitor not running
Daemon: zenvsphere not running
Daemon: zenucsevents not running

However, according to the status in Docker and in the Resource Manager all daemons for that collector are started and running. I also grabbed a copy of the log from /opt/zenoss/log/zenperfsnmp.log as you mentioned and did not find anything peculiar in the log. However, I did see the following lines in the container log for "zenperfsnmp" in Docker:

 sync.go:85: ERR Failed to publish events caused by: EOF

I attached both logs to this post for reference. To troubleshoot further, I started the "zenperfsnmp" daemon manually while still attached to the zenpersnmp container, I removed the device from Resource Manager and re-added, but I am still facing the same issue.

------------------------------
Kai H
------------------------------


Subject: RE: SNMP Collector Creation Failed
Author: Kai H
Posted: 2018-04-06 12:33

Also, I want to mention that the collector host is in its own resource pool separate from the pool that the Resource Manager host is in at the moment. Does this configuration pose a problem? Should the collector host be a member of the same pool as the host for the Resource Manager or does this matter?

------------------------------
Kai H
------------------------------


Subject: RE: SNMP Collector Creation Failed
Author: Kai H
Posted: 2018-04-06 13:50

**From Previous Post**

Also, I want to mention that the collector host is in its own resource pool separate from the pool that the Resource Manager host is in at the moment. Does this configuration pose a problem? Should the collector host be a member of the same pool as the host for the Resource Manager or does this matter?

**************

**So, I tested the theory listed above and removed the resource pool for the collector and added the collector host to the pool with the Resource Manager and it did not make a difference. The issue is still present. :(

------------------------------
Kai H
------------------------------
2018-04-04 18:12:47,640 INFO zen.ZenModeler: Connecting to localhost:8789
2018-04-04 18:12:47,705 INFO zen.ZenModeler: Connected to the zenhub/0 instance
2018-04-04 18:12:47,706 INFO zen.ZenModeler: Getting configuration from ZenHub...
2018-04-04 18:12:48,080 INFO zen.ZenModeler: Starting modeling in 0 seconds.
2018-04-04 18:12:48,080 INFO zen.ZenModeler: Starting collector loop #001...
2018-04-04 18:12:48,080 INFO zen.ZenModeler: Collecting for device 
2018-04-04 18:12:48,080 INFO zen.ZenModeler: Got 1 devices to be scanned during collector loop #001
2018-04-04 18:12:48,220 INFO zen.ZenModeler: Filled collection slots for 1 of 1 devices during collector loop #001
2018-04-04 18:12:48,221 INFO zen.ZenModeler: skipping WMI-based collection, PySamba zenpack not installed
2018-04-04 18:12:48,221 INFO zen.ZenModeler: Collect on device  for collector loop #001
2018-04-04 18:12:48,561 INFO zen.ZenModeler: No Python plugins found for
2018-04-04 18:12:48,565 INFO zen.ZenModeler: No command plugins found for 
2018-04-04 18:12:48,570 INFO zen.ZenModeler: Skipped collection of 
2018-04-04 18:12:48,570 WARNING zen.ZenModeler: SNMP collector creation failed
2018-04-04 18:12:48,575 INFO zen.ZenModeler: No portscan plugins found for 
2018-04-04 18:12:48,575 INFO zen.ZenModeler: Scan time: 0.49 seconds for collector loop #001
2018-04-04 18:12:48,575 INFO zen.ZenModeler: Scanned 1 of 1 devices during collector loop #001
2018-04-04 18:12:48,576 INFO zen.ZenModeler: Daemon ZenModeler shutting down

To troubleshoot, the delegate host (collector) has been rebuilt several times, restarted all collector daemons, and moved the device to the localhost to test modeling. When the device is on the localhost collector, the device is modeled without issue. When moved back to the remote collector, the error listed above is received.

Also, all ports and firewalls have been checked. I am able to connect to all ports to the Resource manager and vice versa with no issues. I have also confirmed that all traffic is flowing between the collector, the monitored devices, and back to the Resource Manager. Any help would be greatly appreciated! Thank you!


Subject: RE: SNMP Collector Creation Failed
Author: Paul Giordano
Posted: 2018-05-15 15:13

Check to see if you have an alert for "SNMP agent down - no response received" on the device you're trying to model. It seems that the modeler will not attempt SNMP if there's an event like that - I closed the one I had on a device and then re-modeled it, and it worked.

------------------------------
Paul Giordano
Senior Systems Engineer
Zethcon Corporation
------------------------------


< Previous
Getting return code from zenmib
  Next
ZSD 6.1.1 - Openstack Provider View Modelling Error
>