Discovery : Schedule Job's getting cancelled.

[Keerthy S]

Hi All, 

Need your kind inputs for this issues. 

Many of the schedule jobs are getting failed. Shazzam takes more time to execute, for example one schedule job has only 5 IP's configured. but still it take very long time to execute 

Action Taken so far:

1. Increased the maximum time for the schedule job, But still jobs gets failed.

2. Verified with Network team, is there any Network Glitch, They confirmed no issues from their end. 

3. Enabled JSON coding of Shazzam payloads in system properties

4. I though due to over lap of schedule jobs may cause the issue, so i have ran the schedule job separately where other jobs are inactive. But still same result(failed) 

Question:

1. Can you please guide me how many IP's are recommended to configure in single scheduled job.

2. We do have 50+ jobs scheduled in one MID Server, where few jobs getting started before the previous job gets finish. most of the jobs cancelled due to Maximum time reached. 

3. is MID Server reboot per weekly/monthly is required for better performance.?

4. in our scenario, We do have 50+ schedule jobs as per the requirement, we are running the discovery during night hours, at maximum 8 to 10 hrs we do have. Can you please suggest method to schedule these jobs (or) do i need a other MID Server ? 

Looking towards your inputs. 

Thank You.

Regards,

Keerthy.  

[Michael de Boer]

How long does it take to do a quick discovery of one of the five IP addresses?
When no other jobs are active on the MID server.

Where is the MID server placed?
Does discovery needs to go through a lot of network devices (also firewalls) to discover the target?

How many ports do you scan with shazzam?
Out of the box it checks around 9 ports.

Regards,

Michael

[Keerthy S]

Hi @Michael de Boer 

Thanks for your kind response. 

1. Today again i have re-ran the job, other discovery jobs are inactive. Job ran successfully. 

2. MID Server is installed -> On-Prem

3. I have checked the inputs queue payload, we are running 14 Ports as below

•   TCP:    HTTP, HTTPS, WMI, WINRM, WBEM, SSH, VMAPP, VMAPP, VMAPP, WINRM_SSL
•   UDP:   WIN , DNS, SLP, SNMP

4. Since we do have 50+ jobs in one server, can you please guide me how to schedule the discovery job without overlapping(few jobs runs for 4 Hrs... few Jobs runs for 1 Hrs.. etc) . (or) do we need new MID Server ?

Thanks, 

Keerthy. 

[Michael de Boer]

I suggest to add one or more MID servers (on another server) and create a load balace MID cluster.
https://docs.servicenow.com/bundle/sandiego-servicenow-platform/page/product/mid-server/task/t_Confi...

You can also check what the performance is of the server during discovery.
When the memory usage isn't high during discovery, you can change the memory usage of the MID in the config. Default it only uses 1024 MB.
When memory and cpu usage are both low, you can also change the number of jobs the MID server can handle simultaneously. Default is 25.
Changing this will speed up the discovery.

To prevent overlapping discovery schedule, it's possible to set the discovery schedule "run" field to "After Discovery" and select another discovery schedule.
This will start the schedule when the previous schedule is finished.

Regards,

Michael