I do not know if it's a "bug" or a "feature" so I have to ask to more experienced people.
The scenario is the one where you have to automatize the installation of the TTU on a ton of client.
Might be the client (pc) have already a TTU's version istalled or not.
Might be on the pc there is a process using the client or not.
So in order to make a robust script I was doing some test/experiment and I have found one issue using the script uninstall_TTU.vbs.
To show the "problem" :
A)On a machine I install :
-Shared ICU Libraries for Teradata
-Teradata GSS Client nt-i386
-Teradata CLIv2 versione
-ODBC Driver for Teradata
-OLE DB Provider for Teradata
B) on a new DOS Shell window I run a script that open an ODBC connection and execute a query every 1 second
C) on a new DOS Shell window I run a script that open an OLEDB connection and execute a query every 1 second
At this point there are two process using the client (dll)
D) on a new DOS shell window I run "cscript uninstall_TTU.vbs ALL /PRIORTOVERSION:99.00.00.00" and wait the completation
Since there are at least two process using the ODBC & OLEDDB dll's my expectation was that the uninstall_TTU.vbs *fails* or return an exit code to perform a system restart to complete the unistallation.
Why the uninstall_TTU.vbs do not return a specific exit code to force the system restart? It's expected by design?
After the above step D you can istall the same tools/product of step A *but* at different version level.
So on the system you have some process using the same utlity/tool but a different version level, it's safe?