Its use can cause query times to be longer on devices with badly designed SNMP stacks (insisting on sending the maximum number of entries per query regardless of how many valid entries there are, for example)

On 2018-01-31 15:45:09, Simon Mousey Smith <simonsmith5521@gmail.com> wrote:

Will it make an improvement if its switched on?

As I was saying before it wasn’t set anywhere

I only added it to my config.php and set as TRUE, I didn’t change anything anywhere

And my polling times increased instead of decreased :( 

Simon

On 31 Jan 2018, at 15:42, Adam Armstrong <adama@observium.org> wrote:

Hmm. No, this has been default to off for quite some time. Perhaps we should default it to on now.

adam.

On 2018-01-31 15:40:20, Adam Armstrong <adama@observium.org> wrote:

Hah, I should have noticed there was no -Cr set on those queries.

BTW don't set things in config use the web config stuff. Did you remove that setting from the config at some point (or did someone unset it from the web config?)

Actually, I think we should be shipping with it enabled as default these days, let me see if setting it to FALSE in defaults is a recent mistake... :D

adam.

On 2018-01-31 15:34:11, Simon Mousey Smith <simonsmith5521@gmail.com> wrote:

Hi,

Its ok I managed to figure it out!

I was reading something last nite somewhere about trying to improve the poller times to get it under the 300s, as mines currently polling out at 334s

And I put in my config.php - $config['snmp']['max-rep'] = TRUE;

The minute I removed this, its back to 9s :) 

Thank you anyways 

And I will remember screenshots next time!

Regards

Simon

On 31 Jan 2018, at 15:30, Adam Armstrong <adama@observium.org> wrote:

I'm not sure I believe that it used to take 9 seconds. (this is why you need to provide screenshots to support any information you give)

There's > 5.5 seconds of ports snmp queries, 2.5 seconds of failed ospf queries, and 3.9 seconds of fdb snmp queries in that log alone.  That is ~12 seconds by itself. 

Why is your device now responding more slowly to queries?

adam.

On 2018-01-31 15:17:45, Simon Mousey Smith <simonsmith5521@gmail.com> wrote:

Hi All

Im trying to work out why 1 of our switches has all of a sudden decided to now take 28s to poll, even tho using it for months its only taken 9s?

From the 'tab=perf' of the device, its showing ‘ports’,’fdb-table’,’netstats’,’ospf’ are showng as really high, 9sec each really?

I have restarted the equipment incase it was maybe that, same issue

Ive upgraded to the latest firmware too but still the same issue?

Any other ideas where I can look?

Output of  /opt/observium/poller.php -h 18 -r -dd      https://pastebin.com/Bdpm3abQ

HP Procurve 1810G - 8GE, P2.22

Regards

Simon

_______________________________________________
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