I have an installation with around a dozen workers configured from the master node. When updating, the worker update process runs fine, until it comes across a machine (cryo16 in this case) which is temporarily off-line. At this point, the following error appears in the console:
Updating worker cryo16
Traceback (most recent call last):
File "", line 24, in
socket.gaierror: [Errno -2] Name or service not known
There is then no attempt to update the remaining workers. In this installation, this doesn't cause a problem, as the worker software is installed in a shared location as suggested in the installation instructions, but could be an issue for other configurations.