When running multiple julia_pods in the same project
(typically via a script that launches them ~at the same time)
I've run into two issues that this fixes:
RUNIDs are not unique enough
(used to be based on datetime with second resolution)
the same secret was being created and deleted, which often
resulted in 'secret not found' errors
This fixes the RUNID uniqueness issue by appending a random number,
and lets secret creation failure not be fatal.
There might be better ways of ensuring secret lifecycles don't
clobber each other. Thoughts @omus?
When running multiple
julia_pod
s in the same project (typically via a script that launches them ~at the same time) I've run into two issues that this fixes:This fixes the RUNID uniqueness issue by appending a random number, and lets secret creation failure not be fatal.
There might be better ways of ensuring secret lifecycles don't clobber each other. Thoughts @omus?