Polling took longer than 5mins

Polling is taking too long due to devices which have a lot of ports.
For example I have a 48-port Cisco SG switch, and it’s taking up to 500 more secs. for that certain device only.

Do you have any idea how could I enhance the polling performance of LibreNMS?

Thank you!

https://docs.librenms.org/Support/Performance/

Specifically selected port polling.

Hi @murrant ,

Already did selected port polling but still no help.
Any other suggestion?

What does your general network performance between LibreNMS and the device look like?
Is there any bottleneck?

@Jarod2801

Network performance is fine. So far both LibreNMS server and the network equipments are in the same switch.

@rysantiago how does the CPU and RAM of the switch look like, any issues there?

@Jarod2801 none as well sir. :slight_smile:

Hello

Is your switch running a fairly recent OS ? If not, I would suggest to upgrade, as SNMP performance and behaviour was pretty bad in older versions.

What is the CPU and Storage performance of your LibreNMS server ? Any IO issues ?

Bye

@murrant @Kevin_Krumm @Jarod2801 @PipoCanaja

I’m trying to look at the poller performance and seems like the reason of poller issue is because of the 48-port of my switch. Though selective port polling is enabled. But I think it’s not enough.

Do you guys have any suggestion to reduce the time for polling per port? Maybe there’s a way to ignore the non-active ports?

Hi,

48 ports on this kind of switch should be done in less than 30 secondes all included. So there is definitly another issue. We have a couple of those SG switches as well, and we have 1 minute polling running stable.
Have a look at the versions and check your LibreNMS server performance to see where the issue is.

PipoCanaja

May I know the specs you have on your LibreNMS server? Is there any minimum requirements for this?

Anyways, this is still on testing mode.

6 vCPUs Xeon E5-2699 @2.3Ghz on ESXi with 6Gb of RAM.
RRDCached is installed
MySQL is tuned (makes a huge difference in our setup, impossible to meet the 1minute without).

RRDCached is also installed in my server.
I haven’t tried MySQL tuning, but may be I should try that one also.

By the way, this is the output during ./validate.php

root@superinfra:/opt/librenms# ./validate.php

Component Version
LibreNMS 1.43-27-g4d99242
DB Schema 263
PHP 7.0.30-0ubuntu0.16.04.1
MySQL 10.0.34-MariaDB-0ubuntu0.16.04.1
RRDTool 1.5.5
SNMP NET-SNMP 5.7.3

====================================

[OK] Composer Version: 1.7.2
[OK] Dependencies up-to-date.
[OK] Database connection successful
[OK] Database schema correct
[FAIL] Some devices have not completed their polling run in 5 minutes, this will create gaps in data.
[FIX] Check your poll log and see: http://docs.librenms.org/Support/Performance/
Devices:
192.168.25.179

@PipoCanaja

Can you please check if I do it right?
Here’s a copy of my screenshot. Thanks!

THANKS @PipoCanaja!!! It works!!!

Thank you for giving me idea. I’m sure this thread could help others as well.

1 Like

Just to add to my own experiences, SG switches are very slow CPU wise, out of everything on our network they poll the slowest.

They’ve made some leeway in speed with SNMP with the latest firmware. I’d recommend upgrading it if you have not already.

@mayheart oh thanks! Maybe I’ll try it also.

Thanks for sharing your knowledge! I really appreciate it.