-
Bug
-
Resolution: Won't Fix
-
Major
-
None
-
A Hudson selenium grid with at leas one JNLP slave on Windows.
Consider the following situation:
Have one Hudson (on Linux) and several Windows slaves (JNLP via Windows service) to form a Selenium grid.
In order to upgrade some plugin or Hudson itself on Hudson you need to shut it down.
When Hudson is restarted all Windows JNLP slave are back, but unfortunately none of them are connected to the Selenium grid.
You need to restart all the Hudson services to have your grid back.
When the connection breaks the launched selenium RC-s remain in running state.
When the connection restored no new RS are lanched, however the old running RC-s won't reconnect.
Either we should kill/restart RC-s or somehow we need to improve the RC to be able to reconnect to the grid, but that's another project.
- is duplicated by
-
JENKINS-5801 Restarting Hudson kills remote RC
- Resolved