Telegraf 1.22.0 / GoSNMP.Conn is missing

[bondskin]

Relevant telegraf.conf

[[inputs.snmp.table]]
    oid = "IF-MIB::ifTable"
    [[inputs.snmp.table.field]]
      is_tag = true
      oid = "IF-MIB::ifDescr"

Logs from Telegraf

Mar 25 10:46:00 raspberrypidns telegraf[16226]: 2022-03-25T09:46:00Z E! [inputs.snmp] Error in plugin: agent dreamer.ip: gathering table unifiIfTable: performing bulk walk for field unifiIfName: &GoSNMP.Conn is missing. Provide a connection or use Connect()
Mar 25 10:46:00 raspberrypidns telegraf[16226]: 2022-03-25T09:46:00Z E! [inputs.snmp] Error in plugin: agent dreamer.ip: gathering table hrProcessorTable: performing bulk walk for field hrProcessorFrwID: &GoSNMP.Conn is missing. Provide a connection or use Connect()
Mar 25 10:46:00 raspberrypidns telegraf[16226]: 2022-03-25T09:46:00Z E! [inputs.snmp] Error in plugin: agent dreamer.ip: gathering table laTable: performing bulk walk for field laNames: &GoSNMP.Conn is missing. Provide a connection or use Connect()
Mar 25 10:46:00 raspberrypidns telegraf[16226]: 2022-03-25T09:46:00Z E! [inputs.snmp] Error in plugin: agent area51.ip: performing get on field sysName: &GoSNMP.Conn is missing. Provide a connection or use Connect()

System info

Telegraf, 1.22.0, Raspberry Pi 4

Docker

No response

Steps to reproduce

...

Expected behavior

no errors and values are correctly captured.

Actual behavior

For some devices, it produces errors and no SNMP values are being captured

Additional info

no issue with telegraf_1.20.4-1_armhf.deb

[Hipska]

This seems like a duplicate of #10554

[sjwang90]

Let's close this to avoid duplicated. @bondskin - please check out #10554 and comment anything you would like to add on that issue.

[Hipska]

@bondskin I just came to some code that might have introduced your problem: #10268

I think we still need to do that reconnect in some cases. Like in your case for example.

If I'm correct, latest working version was 1.20.4? Could you confirm that?

[bondskin]

Hi @Hipska , yes exactly. 1.20.4 was the last good version in regards to this issue.

[Hipska]

Okay, and what are your settings for retries and is it an TCP device?

[bondskin]

 # Polling interval
  interval = "60s"
  # Timeout for each SNMP query.
  timeout = "10s"
  # Number of retries to attempt within timeout.
  retries = 3
  # SNMP version
  version = 2

and yes, TCP

[Hipska]

Okay, I was suspecting that. I think I can fix it.

[srebhan]

@bondskin can you please confirm that the linked PR fixes your issue!?

[bondskin]

Hi @srebhan , this is looking exceptionally great!

it took 1-2 min until the hostnames of the SNMP clients were correctly resolved after restarting the new Telegraf (PR) instance

May 24 20:06:01 raspberrypidns telegraf[13104]: 2022-05-24T18:06:01Z E! [inputs.snmp] Error in plugin: agent millenniumfalcon.ip: setting up connection: error establishing con
nection to host: dial udp: lookup millenniumfalcon.ip on 127.0.0.1:53: no such host
May 24 20:06:01 raspberrypidns telegraf[13104]: 2022-05-24T18:06:01Z E! [inputs.snmp] Error in plugin: agent flash.ip: setting up connection: error establishing connection to
host: dial udp: lookup flash.ip on 127.0.0.1:53: no such host
May 24 20:06:01 raspberrypidns telegraf[13104]: 2022-05-24T18:06:01Z E! [inputs.snmp] Error in plugin: agent area51.ip: setting up connection: error establishing connection to
host: dial udp: lookup area51.ip on 127.0.0.1:53: no such host
May 24 20:07:01 raspberrypidns telegraf[13104]: 2022-05-24T18:07:01Z E! [inputs.snmp] Error in plugin: agent millenniumfalcon.ip: reconnecting: error establishing connection t
o host: dial udp: lookup millenniumfalcon.ip on 127.0.0.1:53: no such host
May 24 20:07:01 raspberrypidns telegraf[13104]: 2022-05-24T18:07:01Z E! [inputs.snmp] Error in plugin: agent flash.ip: reconnecting: error establishing connection to host: dia
l udp: lookup flash.ip on 127.0.0.1:53: no such host

but then everything is working fine, no more errors and all clients are delivering results

[Hipska]

Hi @bondskin, would you be able to test the builds from #11197 as well? (Or the latest nightly builds) They did a minor change on my code and I just wanted to make sure it still works for you.. (Since we could not reproduce your issues)

[bondskin]

Hi @Hipska , thanks for asking, much appreciated.
Still looking good.