Specifying both host name and CIDR mask is invalid
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| postgresql (Juju Charms Collection) |
New
|
Undecided
|
Unassigned | ||
Bug Description
I'm getting the following when deploying to precise / postgresql 9.1 and forming a relationship between services:
2012-05-11 18:21:17,903 unit:offspring-
2012-05-11 18:21:17,998 unit:offspring-
2012-05-11 18:21:20,186 unit:postgresql/0: hook.output INFO: * The PostgreSQL server failed to start. Please check the log output:
2012-05-11 23:21:19 UTC LOG: specifying both host name and CIDR mask is invalid: "ip-10-
2012-05-11 23:21:19 UTC CONTEXT: line 100 of configuration file "/etc/postgresq
2012-05-11 23:21:19 UTC FATAL: could not load pg_hba.conf
...fail!
This is very similar to #918825 though rather than "/0" it's "/32" which I believe was introduced in the last commit (revno 21)
| Mark Mims (mark-mims) wrote | #1 |
| Reprepro Bot (reprepro) wrote | #2 |
| Timothy R. Chavez (timrchavez) wrote | #3 |
dang... finally I see what's going on here. The interface used to pass the explicit IP address... now it's using `relation-get private-address` to get the client's name.
The charm is getting inconsistent responses to `relation-get private-address` across providers... some give the dns name, some the ip address.
The fix is to make sure it's an actual address before adding it into pg_hba.conf. I'll test that theory across different providers.