snmp configuration file can not poll my switch about snmp_exporter HOT 12 CLOSED

I got the same error message from the snmp exporter when talking to a switch and ended up here after a little searching. Figured I'd add my findings in case somebody else ends up here.

I played about with snmpwalk in the command line, stepping up through the protocol levels (I'm new to SNMP). Note, when it fails, the response message also indicates a timeout.

I ended up with snmp v3 with the security level and user name taken from the switch management interface info. snmpwalk then returned the desired info.
I applied the same parameters to the snmp.yml file, started up the snmp_exporter and it's working ok now.

snmpwalk -v1 -c public 172.ip.add.ress
Timeout: No Response from 172.ip.add.ress

snmpwalk -v2c -c public 172.ip.add.ress
Timeout: No Response from 172.ip.add.ress

snmpwalk -v3 -lnoAuthNoPriv -u guest 172.ip.add.ress
SNMPv2-MIB::sysDescr.0 = STRING: M4300-8X8F ProSAFE 8-port 10GBASE-T and 8-port 10G SFP+, 12.0.2.10, 1.0.0.8
....

Added the following to the snmp.yml file:

default:
  version: 3
  auth:
    username: guest
    security_level: noAuthNoPriv
  walk:
  -....

from snmp_exporter.

This is likely due to timeouts. SNMP devices can be very slow. Have you tested using a utility like curl against the SNMP exporter URL to see if it times out?

The SNMP exporter has a hard-coded timeout of 60 seconds. How long does it take to SNMP walk the device?

from snmp_exporter.

A full walk will (almost) always take longer than the few snmpbulkgets snmp_exporter does by default. That being said, you still want to look at the error it gives you via http.

In general, please never ever make screen shots of text. Past the text in here, instead. Everything else makes it very hard to read, and help you...

from snmp_exporter.

SuperQ how can i make that curl against the SNMP exporter URL to see if it times out?
curl -X GET http://127.0.0.1:9116/
dann 404 page not found !

from snmp_exporter.

whenn i browser http
http://172.21.10.117:9116/snmp?module=default&target=172.16.10.1
give me the this error
An error has occurred during metrics gathering:

error collecting metric Desc{fqName: "snmp_error", help: "Error scraping target", constLabels: {}, variableLabels: []}: Error walking target 172.16.10.1: Request timeout (after 3 retries)

from snmp_exporter.

Yes, it seems like your SNMP device is very slow to respond. The timeouts are currently hard-coded, and you'll have to edit them yourself and re-build the snmp_exporter.

See: https://github.com/prometheus/snmp_exporter/blob/master/collector.go#L42

from snmp_exporter.

i test 6 switches and the same result .....always the status is down
i change the time 8 m and never always offline scrap_interval 8 m
scrap_evaluation 8
never changed do you think is there error with configuration file
with nodes ---ok
with snmp ---not ok

from snmp_exporter.

Changing the scrape_interval will not affect the internal timeouts in the exporter. The problem is the snmp_exporter is unable to communicate with your switches, or they take too long to respond.

from snmp_exporter.

SuperQ thank you ....how can i rebuld the file where can i find collector.go
i am beginner can you help me and say how can i build it agein or where i find the correct information _

from snmp_exporter.

i rebuld it still did not work

from snmp_exporter.

now the switch all up but did not show information .

from snmp_exporter.

Hi, this is a problem with your procurve switch. The IP (VLAN) of the request is another one than the reply address.
The following command will solve your problem:
snmp-server response-source yourIP

from snmp_exporter.