Dear Spencer,

Our storage is pretty fast ... i doubt it is related with disk IO issue.

Looking at observium.log i am seeing this

[2016/03/23 10:03:44 +0800] poller.php(13026): /opt/observium/poller.php: hostname-xxx - 1 devices polled in 118.3 secs
[2016/03/23 10:03:45 +0800] poller.php(15743): /opt/observium/poller.php: hostname-xxx - 1 devices polled in 92.80 secs

Is there any error logs which would indicate that poller unable to probe the device due to unable to reach the network?

Thanks!

Regards,
Edy

On 3/23/2016 9:58 AM, Spencer Ryan wrote:
One of our instances polls just over 200 devices (mostly in the same site) with 16 workers and completes in normally around 80 seconds.

[2016/03/22 21:56:28 -0400] poller-wrapper.py(21016): /opt/observium/poller-wrapper.py: polled 202 devices in 86 seconds with 16 workers

The VM runs on a Pure Storage SSD array though. Maybe try turning your pollers down? In the logs folder obvervium.log should tell you how long it's taking to poll all your devices, what do those average?


Spencer Ryan
| Senior Systems Administrator | sryan@arbor.net
Arbor Networks
+1.734.794.5033 (d) | +1.734.846.2053 (m)

On Tue, Mar 22, 2016 at 9:53 PM, Adam Armstrong <adama@memetic.org> wrote:
Are you sure having 24 threads is a good idea?

Gaps like these are usually caused by network issues, or poller processes getting stacked up and interfering with one another.

You can try using the new 'experimental' ports poller mode : 

$config['enable_ports_separate_walk']   = 1;

You'll be able to see on the poller information screen if it's had a positive/negative effect.

adam.

On 23/03/2016 01:50:43, Pui Edylie <email@edylie.net> wrote:

33  */6   * * *   root    /opt/observium/discovery.php -h all >> /dev/null 2>&1
*/5 *      * * *   root    /opt/observium/discovery.php -h new >> /dev/null 2>&1
*/5 *      * * *   root    /opt/observium/poller-wrapper.py 24 >> /dev/null 2>&1


On 3/23/2016 9:46 AM, Adam Armstrong wrote:
show me : 

cat /etc/cron.d/observium

adam.

On 23/03/2016 01:37:31, Pui Edylie <email@edylie.net> wrote:

Dear Members,

We are running latest commercial version and currently we have about 260 devices with 24 probers.

We are getting graph that is breaking up but the link is ok.



Anyone has any idea?

Thanks!

Regards,
Edy



_______________________________________________
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