~lutostag/juju-wait:master

Branch merges

Related source package recipes

1 recipe using this branch. (?)

Create packaging recipe

Related snap packages

Related charm recipes

866bd05... by Greg Lutostanski on 2016-11-10

workaround for cookie failure

783cc92... by William Grant on 2016-11-10

Parallelise slow leadership check.

get_is_leader "juju run"s on every unit, so it's pretty horribly slow
but also trivially parallelisable.

688e10b... by Stuart Bishop on 2016-10-03

Migrate to git, bump version number

5b027bd... by Stuart Bishop on 2016-09-21

Bump version numbers

20d4ba0... by Stuart Bishop on 2016-09-21

Add -m and --model as aliases to existing -e flag, and delint

228c7b9... by Stuart Bishop on 2016-09-21

[lutostag,r=stub] Extend -e argument to support Juju 2.x

49309d3... by Greg Lutostanski on 2016-09-20

set JUJU_MODEL to allow setting specific model for 2.x, works with same cmdline parameters

279d599... by Stuart Bishop on 2016-09-12

Version bump to 2.3.8

15c9f0e... by Stuart Bishop on 2016-09-12

[fo0bar,r=stub] Ignore update-status hook, check leaders after IDLE_CONFIRMATION loop

7540751... by Ryan Finnie on 2016-09-12

Ignore update-status hook, check leaders after IDLE_CONFIRMATION loop

These two changes are intended to reduce wait time on large
environments.

~Recent juju (1.24+?) regularly runs housekeeping hooks, the most common
of which is update-status. On a large environment, there is almost
always an update-status run in the last 15 seconds, but juju-wait should
not block on this.

Run the leaders block after the IDLE_CONFIRMATION block, as checking
leadership on an environment with many services takes a significant
amount of time, and is the least likely to set ready = False and begin
the whole process again. Running this last will make it so it's
(usually) only ever run once, at the very end.