On March 31, 2015, 7:47 p.m., Simo Sorce wrote:
If instances in NASCENT or ERROR are guarnteed to not cause a failure if first_instance is set to true, this looks good. Not sure what kind of error state you'll get in if that is not the case, but I guess it will be ugly either way.
In the current implementation of the initdb, it's a slow (3-5s) no-op if it is called after the system is already initialized, but I didn't want to count on that always being true, hence skipping it if other instances already exist.
- Stephen
----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviewboard-fedoraserver.rhcloud.com/r/154/#review417 -----------------------------------------------------------
On March 31, 2015, 7:24 p.m., Stephen Gallagher wrote:
This is an automatically generated e-mail. To reply, visit: http://reviewboard-fedoraserver.rhcloud.com/r/154/
(Updated March 31, 2015, 7:24 p.m.)
Review request for RoleKit Mailing List, Miloslav Trmac, Stephen Gallagher, Simo Sorce, and Thomas Woerner.
Bugs: 43 https://fedorahosted.org/rolekit/ticket/43
Repository: rolekit
Description
An instance that failed to deploy should not count as a pre- existing instance. This would prevent the database from being initialized properly.
Diffs
config/roles/databaseserver/role.py e2133535f4d8e59b0c1c27f8d6e171f8910d4d12
Diff: http://reviewboard-fedoraserver.rhcloud.com/r/154/diff/
Testing
Created a first instance in error state (no database name, so nothing happened on disk). Created a second instance with correct input. Success.
Thanks,
Stephen Gallagher