Discovery not running consistently

Tom157 · ‎08-09-2011

Hello,

When I kicked off discovery and show status is active. But looks like it's hung and not doing anything. I found out that when i restart the mid server in service-now. Then the job is running, serveral hours later the systom is back again and I need to keep on restarting the mid server for the discover to run.

Anyone have any ideas?

TIA
Tom

tony_fugere · ‎08-09-2011

Is it too large of a job? Is it a Windows MID server? If so, are you overcommitting the TCP stack? Windows has a limitation that could flood it to a halt. If so, think of splitting up your schedule into smaller subnets or standing up another MID to share the load.

OR

Does the MID server hang due to unresponsive probes? Is the problem with the devices themselves?

It should eventually cancel itself. What is your max run time on the Schedule?

tim_broberg · ‎01-22-2015

Questions to sort out what "Hung and not doing anything" is all about:

If you look at discovery status, how many operations are started and how many are completed?
1. 0 / 0 - The job isn't even starting, so the mid server never even gets any work to do. Something is going on with this instance.
2. 1 / 0 - Shazzam is stuck. Perhaps this is a *huge* subnet you're scanning? Wait longer.
3. (several) / (1 or just a few) - Shazzam is completing, but classification is getting nowhere. The mid server and ECC queue are functional, but classification is getting nowhere.
4. Many / (many - a few) - Individual exploration probes are stuck.
If you look at the ecc queue and filter away the processed ones, what's left?
1. Lots of ready - The mid server isn't picking anything up
2. Lots of processing - The mid server is working on a bunch of stuff.
  1. How old are the oldest probes stuck in processing state?
  2. Are they all of a similar type? That could point to a problem with a particular target type / credential.
  3. If the mid server just seems to be stuck, doing a thread dump and then looking at the wrapper log can give some ideas what is going on.
3. Lots of error - Serious weirdness. Check the log for errors.
If you look at the mid server page, is the mid server up?
Anything interesting in the discovery or mid logs?
1. Out of memory? Try adding memory or reducing max thread count.
2. Timeouts? Do they have anything in common? Check for network issues.

Hope this helps narrow down the hang.

Interested to see what you find,

- Tim.

sahityamalipedd · ‎10-19-2016

(0 / 0 - The job isn't even starting, so the mid server never even gets any work to do. Something is going on with this instance.) i have the same issue what is the problem with my instance?? why my discovery is being struck ?? before it was working .doug.schulze

doug_schulze · ‎11-08-2016

If you have done your due diligence in troubleshooting the midserver, upto and including re-deploying , Id have to humbly suggest talking to support..