ArcGIS Enterprise 11.1: operationalhealth.py bug?

[dky]

Hi,

Environment: ArcGIS Enterprise 11.1 (GIS/Hosting/GeoEvent all federated with Portal) on Windows

Just wanted to report a possible bug:

While troubleshooting another problem, I stumbled across this neat little tool called "operationalhealth.py" in <Portal Install>\tools\operationalhealth. But it appears to be confused if the environment includes other federated servers/components.

Specifically: ArcGIS Enterprise 11.1 running with a federated GeoEvent server -- operationalhealth.py seems to believe we have a managed relational datastore and tile cache registered on the geoevent server. At the moment, there are no datastores registered with the geoevent server.

The error messages are, 

2023-10-24 05:25:07,922 [INFO]  OH31: All data stores registered to https://geoevent-server/geoevent are validating.
2023-10-24 05:25:08,308 [ERROR]  OH36: The relational ArcGIS Data Store component is registered with the site at https://geoevent-server/geoevent, which is not the hosting site for the organization.
2023-10-24 05:25:08,368 [ERROR]  OH36: The tileCache ArcGIS Data Store component is registered with the site at https://geoevent-server/geoevent, which is not the hosting site for the organization.
2023-10-24 05:25:08,426 [CRITICAL]  Traceback (most recent call last):
  File "C:\Program Files\ArcGIS\Portal\tools\operationalhealth\operationalhealth.py", line 1388, in <module>
    opHealth.checkNonHostingArcGISDataStore()
  File "C:\Program Files\ArcGIS\Portal\tools\operationalhealth\operationalhealth.py", line 657, in checkNonHostingArcGISDataStore
    if not server['isHosted'] and 'AGSDataStore' in item['path']:
TypeError: 'NoneType' object is not subscriptable

Hope this helps!

Thanks,

Derek

P.S. Thanks to the individual/s who created (and shipped) this tool!

[ArchitSrivastava]

Hello @dky ,

I agree that this script is very useful. I think that this is a BUG as well. I tried the same with my environment which as follows

• ArcGIS Enterprise 11.1 (GIS/Hosting/Raster Analytics (separate) and Datastore (Relational, Tile Cache and Spatiotemporal (all on separate machines)) all federated with Portal and Geoevent Not federated with Portal) on Windows

In my case it seems to think the following:

• The tileCache ArcGIS Data Store component is registered with the Image Server
• The spatioTemporal ArcGIS Data Store component is registered with Image Server

As shown below:

• Date Time [ERROR] OH36: The tileCache ArcGIS Data Store component is registered with the site at https://webadaptor.domain.com/imgserver11, which is not the hosting site for the organization.
• Date Time [ERROR] OH36: The spatioTemporal ArcGIS Data Store component is registered with the site at https://webadaptor.domain.com/imgserver11, which is not the hosting site for the organization.
• Date Time [CRITICAL] Traceback (most recent call last):
  • File "C:\Program Files\ArcGIS\Portal\tools\operationalhealth\operationalhealth.py", line 1388, in <module>opHealth.checkNonHostingArcGISDataStore()
  • File "C:\Program Files\ArcGIS\Portal\tools\operationalhealth\operationalhealth.py", line 657, in checkNonHostingArcGISDataStore
  • if not server['isHosted'] and 'AGSDataStore' in item['path']:
    TypeError: 'NoneType' object is not subscriptable

However, I noticed something that we both have common error "[ERROR] OH36" which seems to be the cause of the problem. As per the documentation (shared below) "[ERROR] OH36" should not even be checked, the checks are only till "OH35" which seems to be the problem altogether

• Scan your organization for operational health issues—Portal for ArcGIS | Documentation for ArcGIS En...

I will run some more tests and will share the details with you soon.

Hope it helps!

-Archit

[ChristopherPawlyszyn]

This will be fixed in the upcoming 11.2 release. Here's the defect for reference:

BUG-000158207 Operational health script throws exception if machines do not exist for ArcGIS Data Stores like graph and object stores on non-hosting site

[JTessier]

@ChristopherPawlyszyn We hit the same issue with 11.1, is there a new version we can get for 11.1 or can we pull it from an 11.2 install?

[mabdraboh]

Hello everyone,

I have encountered the same issue in the last release 11.2 in two environments (Staging, and Production). I would appreciate if anyone from Esri team could tell if they reported this in the last release.

[ArchitSrivastava]

Hello @mabdraboh, 

I tested this on 11.2 several types of environment listed below and the results :

• ArcGIS Enterprise Single Machine Deployment  : Works
• ArcGIS Enterprise High Availability (With All ArcGIS Server Roles on individual Multi-machine and ArcGIS DataStore all Roles in High-Availability) : Shows "[ERROR] OH36" (even though all Datastore registrations are correct) but proceeds with the process in comparison to previous version where it used to exit the tool run. 

And I can confirm that the tool works in certain scenarios for 11.2 environment but still completes the run. Hence, I would request if you can share more details about the error perhaps we can try to check what is happening.

@JTessier 

I tested what you mentioned as well. I coped the "operationalhealth" folder from 11.2 to 11.1 environment and it gives the same error mentioned above "[ERROR] OH36"  as mentioned before.

Hope it helps!

-Archit

[mabdraboh]

Sorry for this late reply. I have contacted the technical support and they tested several scenarios some of them did not report the bug (single machine deployment) but with multi machine deployment with several ArcGIS Server role they were able to reproduce the issue, so they reported it as bug in 11.2 release. May this help anybody that encountered the same issue.