9.2.0 hub has stopped working

[Robert E]

I upgraded from 8.5.1 to 9.0.2 then to 9.2.0 Monday 9/9.
The server is running Windows Server 2012 R2 and is both my tunnel server and ump server.
After upgrading to 9.2.0 the data_engine 9.20 probe would not start and I was provided the 9.20HF1 which resolved that issue.


I typically do all my administrative work on the primary server using the Infrastructure Manager application.

Since the upgrade I have been unable to start the primary hub on my primary server via the Infrastructure Manager application.

I can start and make changes to that hub probe via the Admin Console.
I can start all other secondary hub probes via the Infrastructure Manager application.

I have rebooted the server and restarted the service multiple times.

In attempting to troubleshoot this with Broadcom IT I have re-installed all of the Visual C++ versions (2008, 2010, 2013, 2018), provided multiple log and config files, deleting all files/folders from %temp%, and still the issue remains.

I also tried downgrading the hub to 7.97 from the archive and still have the same issue.

Broadcom IT indicates this is not a Broadcom issue as the ucrtbase.dll erroring is not a Broadcom dll file.

Has anyone else run into this issue and can provide a solution or troubleshooting tips?

[David Michel]

Are there any other problems administering the primary hub via IM? 
for example:
will probe configure gui open? 
deploy a probe? 
in the archive select a package > Edit?

------------------------------
Support Engineer
Broadcom
------------------------------

[Robert E]

This is the only issue I have noticed since the upgrade (other than the data_management hotfix).

I don't have any issues opening other probe gui's on the primary hub via IM.
All probes I have tried to deploy to the primary hub from the archive were successful.
The edit package GUI of a random probe from the archive loads without issue.

[Daniel Blanco]

What OS are you running the primary hub on? 

Did you make directory backups before upgrading. You could always just revert back the DB and the \Nimsoft directory if you did backups. Step 1 make a backup of NimsoftSLM db and all UMP and Pri-Hub Nimsoft folders..




------------------------------
Daniel Blanco
Enterprise Tools Architect
Alphaserve Technologies
------------------------------

[Robert E]

Windows Server 2012 R2

I had a full snapshot of the server and database backup but I gave the OK to delete them a couple days after was successful before I noticed this issue.
I had opened multiple probes and hubs since the upgrade just not the primary hub it seemed.

I so have a backup of the hub.cfg from 8.5.1 but that is all that remains.

[Gene Howard]

The problem is probably the OS 2016R2
we have only certified up to 2016
https://docops.ca.com/ca-unified-infrastructure-management/9-0-2/en/installing/product-compatibility/compatibility-matrix

The problem you are having is with the hub GUI only not the hub probe.

You might try applying the latest version of the C++ 2017 files from MS.
https://support.microsoft.com/en-us/help/2977003/the-latest-supported-visual-c-downloads

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------

[Robert E]

I apologize, there was a typo in my second posting of my OS.
The server is running Windows Server 2012 R2 as indicated in my first post.

You are correct, the problem is with the GUI of the hub probe on the primary server.
If I open the GUI of any secondary hub probe it opens without error.

As for the Visual C++ 2017 files, I have already re-installed those.
Here are the Visual C++ currently installed:

[David Michel]

Not seeing the 2017 in your image. 
This is from my 9.2.0 primary hub server

both 2008 & 2017 are needed. 

------------------------------
Support Engineer
Broadcom
------------------------------

[David Michel]

Additional info:
the 2008 vs was downloaded from MS site and installed, because with out it IM console doesn't work. 
the 2017 vs was just deployed via the IM archive.

------------------------------
Support Engineer
Broadcom
------------------------------

[Robert E]

I believe it is included in the 2015-2019, because when I try to install 2017 by itself I receive:

[Gene Howard]

the only thing I can think to suggest would be the following.
1) Uninstall all of the visuall C++ run time libraries versions.
2) uninstall IM and the Soap run time.
3) clear the temp folder
4) reboot the server.
5) disable any antivirus
6) install the IM client from the <Nimsoft>\install\setup
7) deploy the visuall C++ 2010sp1 and vs2017 (both x86 and x64) from the archive to the robot.

And see if this does not resolve the issue.

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------

[Robert E]

I uninstalled all of the Visual C++ versions, Nimsoft Infrastructure Manager, SOAP-runtime-TK3
I cleared the temp folder and restarted the server. 

After rebooting I installed the IM but received several registration errors:

I uninstalled the IM and SOAP again and installed Visual C++ 2008 x86 & x64 and reinstalled the IM.
This time only the nimbus.dll failed to register. 

I restated the server again and launched the IM but it failed to launch.  The error was: "Check if 'nimbus.dll' is properly registered".
I installed Visual C++ 2017 from MS and restarted.

After the restart the IM starts without error, but I still receive the same 'hub has stopped working' error when attempting to launch the GUI.
All other probe GUI's and secondary HUB gui's that I tried open without error.

I installed the vs2010sp1... and vs2017... from the archive and restarted the server again.

I still receive the same 'hub has stopped working' error when attempting to launch the GUI of the primary server.
All other probe GUI's and secondary HUB GUI's that I tried open successfully.

[Gene Howard]

I think the dev team will have to look into the application dump file to find a solution to this at this point.

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------

[Robert E]

Gene,

How would I engage the DEV team?

Thanks,
Robert

[Gene Howard]

this would have to be done through a support case.

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------

[Robert E]

So in working with Broadcom Support/Software Engineering they have asked that I try logging into the primary hub from a secondary hub using the Infrastructure Manager application.

Once ports 48000-48050 had been unblocked (everything runs through tunnel server port) and the secondary hub can now see the primary hub in the Infrastructure Manager application, I tried logging in only to get the login/pass error below.

Anyone have any suggestions on what I need to do to allow login?  This is the same login & password that I use to login on the primary hub.
Broadcom support suggested disabling local IP validation on both the controller and the primary hub, before referring me back to the Broadcom Communities as that didn't work.

[Gene Howard]

UIM names are case sensitive.
Please try administrator instead of Adminsitrator

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------

[Garin Walsh]

And if not that, sometimes it takes time for the security file to be passed around. Each hub authenticates off it's own local copy of security.cfg and sometimes that doesn't get updated. It should be almost the same in size across all systems and inside it in the <setup> stanza, there's a "version" key which should be the same too.

If not, you will probably just have to wait an hour or two more and try again.

[Robert E]

Well don't I feel silly for not trying 'administrator'.

Strangely opening the primary hub gui still fails to open.  
I'll report back to support.

Thanks!

[Robert E]

FYI, issue of opening the primary hub GUI was resolved when support provided me a replacement nimbus.dll