Why Discovery Schedules are hanging when multiple CI IPs addresses are associated with a single device?

[Vicente]

Hello Community.

While running discovery schedules some IPs addresses are hanging and discovery never finishes.
I noticed that the IPs addresses that are hanging in the scanning phase are associated with particular devices.
These devices have multiple IPs addresses configured in its LAN nic.

For example; we have a server with 5 IPs addresses configured in its LAN nic.
Discovery updates the CI record using just one of the IP addresses and it hangs trying to scan the 4 remaining IPs addresses till the configured max run timer in the discovery schedule runs out of time.

I did a test scanning one by one the 5 IPs and the CI record was being successfully updated each time I scanned the IP.
The issue happens when Discovery is scanning all of the IPs together. I could exclude 4 of the IPs addresses so those wouldn't be scanned; however, I don't know what those IPs are used for.

Your comments/suggestions are welcome.

Thanks.

[tim_broberg]

There it is. Looks like you posted it the first time, and I'm just being obtuse.

Ok, so, I think what you're saying is, it only runs all those exploration probes on one of the IPs per discovery?

That is the expected behavior.

1. "Shazzam" Port scan : (One probe for many IPs) Find what IPs have what ports open.
2. Classify: (One probe for each IP) Find out what kind of device is present on all those IPs.
3. Identify: (One probe for each IP) Figure out which CIs all those IPs belong to.
4. Explore: (Many probes / patterns for one IP selected from each CI) Examine one IP for each CI, and collect all the relevant information.

We don't need to explore the device through more than one IP as it reports the same information regardless of how we get in.

Does that answer the question, or am I barking up the wrong tree?
    - Tim.

[Vicente]

Yes, Tim. What you already state makes total sense to me. So, do you suggest that I should exclude the other 4 IP addresses in order to prevent the Discovery schedule keeps hanging each time it runs? 

[tim_broberg]

Ok, so what does hanging look like, on the ecc_queue?

There's an output probe in processing state with no input?

If so, you might want to get a stack trace and see where it's stuck.

If not, then what does the ecc_queue look like for that IP?

[Vicente · Accepted solution]

Regarding the schedules that are timing out, I find out these links useful:

https://community.servicenow.com/community?id=community_question&sys_id=d26bcb21db9cdbc01dcaf3231f96...

https://hi.service-now.com/kb_view.do?sysparm_article=KB0722882

We implemented workaround number 2 and it worked.

Also, ServiceNow say this issue was fixed in Madrid Patch 3 and New York versions; but our instance is running on Madrid patch7a and we were still facing the issue.
https://docs.servicenow.com/bundle/madrid-release-notes/page/release-notes/quality/madrid-patch-3.ht...

Hope this helps.

[Community Alums]

We are also facing the same issue. Discovery scan is getting cancelled. Can I know the workaround that worked for you to solve the issue.  the link provided is from Hi and its not opening.