| From: | Heikki Linnakangas <hlinnakangas(at)vmware(dot)com> |
|---|---|
| To: | Erik Rijkers <er(at)xs4all(dot)nl>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: pg_isready --username seems an empty promise |
| Date: | 2014-05-19 11:22:12 |
| Message-ID: | 5379E964.4010303@vmware.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 05/19/2014 01:37 PM, Erik Rijkers wrote:
> pg_isready has --username:
>
> -U, --username=USERNAME user name to connect as
>
>
> so is replying when given a non-existent user not a bug?
>
> pg_isready --username=xxxx -p 6544
> /tmp:6544 - accepting connections
>
> There is no user xxxx. (PG envvars are removed)
Per the manual page on pg_isready:
> The options --dbname and --username can be used to avoid gratuitous
> error messages in the logs, but are not necessary for proper
> functionality.
The libpq entry on PQpingParams(), which pg_isready uses, has a bit more
details:
> It is not necessary to supply correct user name, password, or
> database name values to obtain the server status; however, if
> incorrect values are provided, the server will log a failed
> connection attempt.
BTW, I find the above PQpingParams() explanation much more clear than
one in pg_isready. I think we should use the same text in both places.
- Heikki
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Christoph Berg | 2014-05-19 11:53:18 | Re: 9.4 beta1 crash on Debian sid/i386 |
| Previous Message | Heikki Linnakangas | 2014-05-19 11:16:15 | Re: cosmetic fixes |