What version of nx-os are you running?5.x and below sucks att SNMP, 6.x is ok bit 7.x actually works./MarkusDen 2 nov. 2017 18:44 skrev "Stef Renders" <stef.renders@cronos.be>:That’s perfectly ‘normal’ for these devices, as it creates virtual interfaces for each vnic/vfc in each blade / server in the fabric, the amount of interfaces to poll add up very quickly
We see the same on 5k platform when you have a few fex’s attached. Both of them averaging a little less than 5 minute poll time.
Also, having 3500 vlans (thus thousands of macs) on them as we do does not help
(touchpad drawings ftw)
Kind regards
From: observium [mailto:observium-bounces@
observium.org ] On Behalf Of Edvinas K
Sent: donderdag 2 november 2017 17:52
To: Observium <observium@observium.org>
Subject: Re: [Observium] some devices takes very long to complete /opt/observium/poller.php process
Sorry, it's the Cisco Nexus 6000 Switch.
Maybe it could be due to the distance to the device ? (the round trip from server to device is - 275ms.)
On Thu, Nov 2, 2017 at 6:49 PM, Adam Armstrong <adama@observium.org> wrote:
It means your device is slow and taking too long to return the requested data.
You've not actually really included any useful information though. What is the device?
adam.
On 2017-11-02 16:34:36, Edvinas K <edvinas.email@gmail.com> wrote:
Hello,
Some devices takes very long to complete /opt/observium/
poller.php process.
For example i took one of them and run:
/opt/observium/poller-wrapper.
py --host hk2net01a -d
/opt/observium/poller-wrapper.
py:568: Warning: Data truncated for column 'process_name' at row 1 cursor.execute(p_query, (pid,ppid,processname,uid,
command,s_time)) INFO: starting the poller at 2017/11/02 17:10:50 with 16 threads
WARNING: DEBUG enabled, each device poller store output to /tmp/observium_poller_id.debug (where id is device_id)
/usr/bin/env php /opt/observium/poller.php -d -h 320 >> /tmp/observium_poller_320.
debug 2>&1 INFO: starting alerter.php for 320
INFO: finished alerter.php for 320
WARNING: worker Thread-1 finished device 320 in 405 seconds
INFO: poller-wrapper.py poller --host hk2net01a processed 1 devices in 405.53 seconds with 16 threads, load average (5min) 1.67
WARNING: the process took more than 5 minutes to finish, you need faster hardware or more threads
INFO: in sequential style processing the elapsed time would have been: 405 seconds
WARNING: device 320 is taking too long: 405 seconds
ERROR: Some devices are taking more than 300 seconds, the script cannot recommend you what to do.
Number of rows updated: 1
DEBUG: /usr/bin/rrdtool update /opt/observium/rrd/poller-
wrapper.rrd N:1:405.531:16 DEBUG: /usr/bin/rrdtool update /opt/observium/rrd/poller-
wrapper_count.rrd N:1
The debug txt shows that it freezes for about 4mins on this line:
"SQL RUNTIME[0.00020313s]
o Caching Oids Used full table ifEntry/ifXEntry snmpwalk.
ifEntry
CMD[/usr/bin/snmpbulkwalk -v2c -c *** -Pu -OQUs -m IF-MIB -M /opt/observium/mibs/rfc:/opt/
observium/mibs/net-snmp 'udp':'hk2net01a':'161' ifEntry] ...................."
and almost after 5mins it resumes. What could be the reason ?
Thanks.
______________________________
_________________ observium mailing list observium@observium.org http://postman.memetic.org/ cgi-bin/mailman/listinfo/ observium
_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/ observium
_______________________________________________
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/ observium
observium mailing list
observium@observium.org
http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
