Address cases when reused qtlicd gets killed

When using QLS CIP in Squish, we stumbled upon the situation when reused qtlicd got killed, and the client which was using it throws the error:

Failed to send data: TCP: Error while sending data: An existing connection was forcibly closed by the remote host.
(status code: 207)

The situation was like this:

• using CIP, the process A indirectly starts qtlicd.exe with: qtlicd.exe --mode on-demand --nostdout
  (qtlicd.exe is a sub-process of the process A)
• the process B also uses CIP and needs qtlicd.exe in the on demand mode
  (CIP will ruse qtlicd.exe from the process A, because that qtlicd.exe is already running)
• process A and its sub-process get killed for some reasons
  (there is no more running qtlicd.exe on the system)
• after some time (a minute or two), the process B will print the error above

It seems that relaying on the shared (reused) qtlicd to not get killed by its parents might create problems in some scenarios. It would be good to address this in some way.