On a minor interactive update, we never prompoted to clear breakpoints
when using deployed-server since the code reads the server id's from
nova, of which there are none.
This modifies the behavior to read the server id's and names from Heat
when nova returns no servers.
Change-Id: I682f6dc66705c9d42b9c2d21f675491ea60c9c3c
Closes-Bug: #1708236
The update abort is not reliable, and could mess up a TripleO stack
beyond repair. Since this has a potential for data loss, I would
suggest this be removed without a deprecation period.
Depends-On: Ieec4f01e38768eafb3df1f06340bdd3e220d30bd
Change-Id: I752e061979d667c1fb2b115c1a7339002e1824d5
Since Iaacc585b9806c08fd5a57c80c10eeb31cb0a8f71 got merged in Heat, we
can get into a state where a failed resource is recreated during update
after clearing a hook. That means that when we poll for events, we need
to check for that additional action and not limit to updates.
Change-Id: I2191a3f3df426eaa91fdd5b00299193a32efe366
Closes-Bug: #1677548
When removing a breakpoint fails because the resource name was not
found, the error would be reported using either the previous server_name
or the client would crash with the error "local variable 'server_name'
referenced before assignment".
Change-Id: I35eb2359dbfc9b678dcd26a33766d8c19a456539
This patch is replacing raw_input() with six.moves.input() because
raw_input is not present in python3+.
Fixing this also makes "tox -epep8" happy when running with python3.
Change-Id: If28f97b479fa018079c7d180541ca29689398a68
We first need to wait for the stack to begin updating before we start
checking if it's complete/failed. If we don't do that we can hit a race
condition between the CLI and Heat where CLI thinks that the update has
finished, but in fact it hasn't even started yet.
Closes-Bug: #1589986
Change-Id: I755fe02155a37698f8c9c885d7b6bea984e21a75
LOG.warn is deprecated. It still used in a few places.
Updated to non-deprecated LOG.warning.
Change-Id: I897fc367da8a8d003cced2177aea680423ec561a
Closes-Bug:#1508442
Thomas Herve