lp://qastaging/~dpb/charms/precise/postgresql/extra-packages

Branch merges

Propose for merging

No branches dependent on this one.

Merged into lp://qastaging/charms/postgresql at revision 36

Mark Mims (community): Approve on 2013-02-15

Diff: 142 lines (+33/-24)

3 files modified

README.markdown (+16/-11)
config.yaml (+4/-0)
hooks/hooks.py (+13/-13)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

35. By David Britton on 2013-02-15

Merging up...

34. By David Britton on 2013-02-14

hba_conf changes do not need full database restarts

33. By David Britton on 2013-02-14

- Fixed Indention typo that overwrote data if multiple units join the relation.
- Optimized Install -- multiple packages with one command
- Addded config option to install extra packages
- Updated README slightly

32. By Marco Ceppi on 2013-02-13

Marco Ceppi 2013-02-12 Added check for ip or hostname

31. By Marco Ceppi on 2013-02-07

Marco Ceppi 2013-02-06 Install postresql-contrib

30. By Marco Ceppi on 2013-02-07

Marco Ceppi 2013-02-06 Since we're working with explicit addresses don't appent /32

29. By Juan L. Negron on 2013-01-15

Merging Sidnei's bag of tricks. MP:143097

28. By Juan L. Negron on 2012-11-28

While putting together a charm with a postgresql relation, I ran into a situation that appears to have been anticipated by a TODO comment in the postgresql charm:

    # XXX: This appears to reset the password every time
    # this might not end well

Specifically, I found that the password was being reset (IIRC by the db-{,admin-}relation-changed hook) after I'd made use of it in my charm, and of course then everything went wrong.

There doesn't seem any particularly nice way to deal with this; I don't think we can rely on the password being cached in the relation, because multiple relations might make use of the same user. But given the existing password handling, it doesn't seem that bad to just cache it in a file readable only by root, and this fixed the problem for me. Of course I'm open to better suggestions. MP: 135128

27. By Juan L. Negron on 2012-11-28

I have a charm that needs the debversion type in PostgreSQL. As far as I can tell, there seems no harm in just trying to install this unconditionally. MP: 135130

26. By Juan L. Negron on 2012-11-28

The postgresql charm includes *-relation-departed hook symlinks, but hooks.py doesn't implement them and exits 1 when they're called. This has the effect that whenever a condition arises such that a *-relation-departed hook is called, the relation goes into an error state.

I don't know if *-relation-departed should be implemented (I had some trouble divining the exact difference between that and *-relation-broken) but it would be better for them to be absent than broken. MP:135132