[DPE-5588] Check against invalid arch #539
Conversation
sinclert-canonical
force-pushed
the
sinclert/5588/arch-check
branch
3 times, most recently
from
125f6ab to
e3cab7b
Compare
sinclert-canonical
requested review from
taurus-forever,
shayancanonical and
carlcsaposs-canonical
There was a problem hiding this comment.
we may want to reconsider our approach here for in-place upgrades/rollbacks
(i.e. if someone refreshes from correct arch to incorrect arch)
at least with refresh v3, this change may prevent rollbacks that we could potentially recover from on the wrong arch, but ensuring rollback safety might be a bit complicated
sinclert-canonical
force-pushed
the
sinclert/5588/arch-check
branch
from
cd91153 to
751a0b4
Compare
sinclert-canonical
force-pushed
the
sinclert/5588/arch-check
branch
from
751a0b4 to
7686ce3
Compare
There was a problem hiding this comment.
still have concerns about rollbacks & how this could makes things worse than if we continue to raise uncaught exception (#539 (review))
everything else looks good!
Not sure what you mean by that. Mind bringing this up during today's Technical Sync? |
This PR introduces the usage of the newly created architecture lib (see #560) into the charm itself.
The shape of the integration tests, as well as the helper used to get the built charm in the architecture that mismatches the one of the host itself, have been adapted from the one Lucas created for PostgreSQL-k8s (see code).