Shazzam probe being stuck

Paul125
Kilo Guru

Hello, we have weekly discovery schedules set up in place and have been working for years. All of sudden these schedules being canceled it self due to time out. after digging further, we found that 'shazzam' probe being sent out but 'processed' state is showing processed but timestamp showing empty and 'updated on' showing discovery end time like below. Has anyone experienced before? Thanks

find_real_file.png

1 ACCEPTED SOLUTION

Paul125
Kilo Guru

Restarting the MID solved my issue.

View solution in original post

22 REPLIES 22

Branden
ServiceNow Employee
ServiceNow Employee

ALMOST sounds like the ECC queue got "stuck". 

 

I'd go into the vanilla ECC list (making sure no other discoveries of ECC use mechanisms aren't active) and filter to anything that's not a good state (probably processing).  Then manually set it to canceled or something.  I've not seen this since H though, to be honest. 

 

If it happens again it's worth a HI ticket for sure. 

Thanks for the reply Brandon. Here is another situation where probe classify is being stuck if the Shazzam probe gets processed on time.

tim_broberg
ServiceNow Employee
ServiceNow Employee

If I recall correctly, the probes will all get marked processed when the discovery gets canceled.

The question is, what's going on that this shazzam is stalling?

Take a look at the log files in agent/logs. The wrapper log will show events that happen to the JVM as a whole, like out of memory and restarts. Usually you look there last, but give that you're seeing strange things in shazzam and classify, I would start there first this time. Then you can look at agent0.log.* to get a blow-by-blow description of what is happening.

    - Tim.

Thanks for the reply Tim. Not only the Shazzam but also WMI: CPU / Memory, Identification probes as well but Shazzam is more like 80%. We have 10 MID servers pretty much all MIDS have the same issue but 4,5 MID servers has more discovery failures and Shazzam issue.