Solved

FME Startup Python Script: Kerberos Ticket Cache not updated


Badge +1

We have a startup python script that will run a kinit process using a keytab file on the server. Unfortunately it appears that even though the kinit was successful and the Kerberos ticket was created correctly, FME is not able to see the newly created ticket in the Kerberos ticket cache. Once the current FME process terminates any new FME processes after wards can see the new ticket. It seems to me the Kerberos ticket cache is read into memory with the initial FME.exe starts and updates are not known about by the FME process. Any idea how make the ticket cache update during the current FME process?

 

Any ideas or work arounds would be helpful! Current work around idea is to have a separate FME process run separately to generate the Kerberos tickets.

 

Thanks in advance!!

icon

Best answer by debbiatsafe 18 July 2019, 21:18

View original

2 replies

Userlevel 3
Badge +17

Hi @justincornell

I'm not personally familiar with kinit processes but I'll do my best to help. How are you trying to access the Kerberos ticket in your workflow?

It does sound like the simplest solution would be to separate your workflow into two parts (eg. using a WorkspaceRunner or something similar).

Badge +1
This was fixed in 2019.2 and the kerberos cache can be modified in the FME.exe process and it will pickup any modifications. Therefore I can have a startup script create a new kerberos ticket and my jdbc database driver will pick it up.

Reply