SNMP doesnt work for some devices

Help
1

Hello,

using Librenms for few days. Got about 40 devices in it, Huawei/HPE.

Some of HPE switches doesnt work, i will describe problem on one of them, rest got same symptoms.

(ips changed)

HPE1 - 192.168.0.1
HPE2 - 192.168.0.2

Both are TOR switches, physically in racks next to eachother.
Identical configuration (except management ip), identical os version. Load/traffic on both of them is ~same.
But HPE2 is not showing any data in librenms.

When HPE2 is deleted from librenms, snmpwalk works like a charm, i can run it whole day without issue.
But when i add HPE2 to Librenms, i get Timeout: No Response from udp:192.168.0.2:161
Also, during pooling, snmpwalk/snmpget is not working against HPE2.

I wrote a script which tries snmpget each 5 seconds, you can see timeouts which happens when librenms starts pooling:
https://p.libren.ms/view/004792ad

Pooling output:
https://p.libren.ms/view/dbea23d9

validate.php:
https://p.libren.ms/view/ff7a7e62

this one is not clean, but 35 out of 40 devices works just fine, so i dont think it will cause this issue.

Any ideas how to troubleshoot this?
Thank you!

2

What snmp version are using?

Also have you checked the performance docs for snmp timeouts?

3

There are many tools you can use to read SNMP from devices, simple free tools/utilities. I would check that first from your workstation. If in Windows try MIB Browser. Very small install, enter the device IP or hostname and the SNMP credentials…and see if it’s responding at all.
Then from the console on your LibreNMS system you can use the linux tool SNMPWALK to see if SNMP is responding to that system. You can at least eliminate what might be a simple SNMP configuration problem on the devices to be monitored.

1 Like
4

Can you tell me which model HP switches give this problem?
Then I can check if the problem occurs even when tired and see if hp are MIBS for this.

5

Hello,
sorry for late answers, i was sick :frowning:

Anyway, few updates:

  1. Definitely not a issue with librenms, i tried to install zabbix and prtg and exactly same problem with same switches, so there is some snmp problem on my switches.

I would be really glad if i can continue to troubleshoot this, but if its out of scope of this forum, feel free to delete this post.

  1. snmp version doesnt matter, v2/v3, same situation
  2. Like i wrote, snmpwalk is ok until its run from any monitoring system, then my snmp on swiitch freeze.
    COnfiguration is ok, other switches have identical configuration, identical os, identical hardware.

@Svenman2 its A5500-48G-4SFP HI Switch , thanks for any help!