monitors for one of the SP's goes "UNKNOWN" frequently

[lravelo]

I'm running this against a Unity 450F on Check_MK Enterprise v1.6.0p22 and frequently the service for CPU utilization and Temperature goes unknown for about 15 minutes or so. This happens basically every hour from what I can see. How can I troubleshoot this to identify what the cause is?

[ImTheLucKyOne]

The temp and cpu checks get their values from the uemcli metrics output. So what you can try first is to run the uemcli manually when the issue occurs:

# uemcli -d HOST -u USER -p PASSWORD /metrics/value/rt -path sp.*.platform.storageProcessorTemperature show -interval 5 -count 1 -output csv
# uemcli -d HOST -u USER -p PASSWORD /metrics/value/rt -path sp.*.cpu.summary.utilization show -interval 5 -count 1 -output csv

Depending on that output while the cmk services are unknown you can investigate further.

[lravelo]

Thanks for your response. This is the output I get when I run these commands:

./uemcli: error while loading shared libraries: libboost_thread-gcc41-mt-1_43.so.1.43.0: cannot open shared object file: No such file or directory

EDIT:

I used the uemcli.sh script instead but it's giving me a permissions issue even though this is the same account that I use to monitor and is working fine:

Operation failed. Error code: 0x6000001 You do not have access to perform the requested operation. (Error Code:0x6000001)

EDIT 2:

For some reason the commands started working again after resetting the password. Next time the alarms come up in checkmk I will test the commands you've mentioned.

[lravelo]

after resetting the password, everything went from bad to worse. Now I can no longer monitor anything. I've deleted everything in the ./omd/sites/<monitoring_site>/.emc and reran uemcli -d HOST -u user -securePassword -sslPolicy store /sys/general show to download the certs, deleted the host from check_mk, deleted the rule for the special agent and nothing. The only thing I get back is results from the emcunity_hostcons check and it comes back as unknown.

EDIT:
Turns out this was related to some character in the new password. It's reporting again. Will check again once the SP's cpu and temp services come back unknown.

[lravelo]

This is the only thing being returned when it happens:

[ImTheLucKyOne]

Well, that looks like it should.

Next step I would do, just to double check, run the agent manually while the UNKNOWN issue occurs and see what it reports back.