Stacked Juniper EX2200 switches services in stale or pending

Troubleshooting
1

stacked Juniper EX2200 switches services in stale or pending they are not coming up .please let me know how to resolve this issue.

2

Juniper switches with many ports need some optimizations wrt their SNMP agent:

https://www.juniper.net/documentation/en_US/junos/topics/task/configuration/snmp-best-practices-device-end-optimizing.html
https://blog.proact.de/2015/05/27/snmp-optimierung-unter-junos/

You can also adjust timing settings on the CheckMK side.

3

How can i adjust the timing settings in check_mk.if we adjust the timing what is the use of it.

4
  • WATO → Host & Service Parameters → Access to Agents → Timing settings for SNMP access
    This rule decides about the number of retries and timeout values for the SNMP access to devices.

You can adjust Response timeout for a single query and Number of retries

Perhaps have a look at the following ruleset as well:

  • WATO → Host & Service Parameters → Access to Agents → Check intervals for SNMP checks
    This rule can be used to customize the check interval of each SNMP based check. With this option it is possible to configure a longer check interval for specific checks, than then normal check interval.

Well, increase timeouts so Checkmk will wait long enough for the checks to finish. :wink:
(But please also check out the links @r.sander posted.)

5

thank you all your help :slightly_smiling_face:
still i am unable to resolve my problem.

6

I have the same problem with an SRX. I set timeout to 60s and polling interval to 5m. I can occasionally get a discovery to run and pull the interfaces, but it will go stale after next round of checks. If I successfully rediscover, it will all go green again.

If I remove all of the network interfaces, it will poll correctly.

Discoveries that fail respond with:
Your request timed out after 110 seconds. This issue may be related to a local configuration problem or a request which works with a too large number of objects. But if you think this issue is a bug, please send a crash report.

7

To get a good feeling what takes so long, it is good to run the discovery on the command line. There you can measure the time needed for a complete discovery or a complete check run.

cmk --debug -vvI <hostname>     (discovery)

or

cmk --debug -vv <hostname>      (check)
8

Running the manual discovery completed successfully (and cleared all of the stale checks at the same time). It took a really long time to complete.

Edit to add:
ran the manual check immediately after and it completed <1s
Ran it again and it’s taking minutes

OK - [snmp] Success, execution time 222.5 sec | execution_time=222.536 user_time=0.180 system_time=0.030 children_user_time=0.000 children_system_time=0.000 cmk_time_snmp=222.321 cmk_time_agent=0.001

9

I turned off network interfaces and it doesn’t go stale. However I have no way to get into services in WATO to fix this as host/services edit always times out :

Check_MK Discovery [](javascript:void(0):wink: CRIT - no unmonitored services found, 10 vanished services (snmp_uptime:1, juniper_mem:2, juniper_fru.fan:2, juniper_fru:2, juniper_cpu_util:2, snmp_info:1), no new host labels, [snmp] keepalive timed out CRIT

Edit to Add - it seems to go in and out of crit for this check.