[Pgpool-hackers] retry of health check

[Fujii Masao]

On Wed, Mar 10, 2010 at 6:21 PM, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> No. If postgres database does not exist, backend returns 'R' instead
> of 'E'.  So pgpool writes 'X' then fails and retries with template1.

Hmm.. in HEAD, 'E' is returned first in that case. Which version of
postgres returns 'R' first? Its behavior might depend on the version.

On second thought, in the version where 'E' is returned first,
nonexistence of 'postgres' database is treated as a success of the
health check. OTOH, in the (old?) version where 'R' is returned
first, that error might be treated as a failure and cause the retry.
This version-related difference is also confusing for me.

> How do you cope with older PostgreSQL version which does not have
> postgres database then?

Yeah, the backward-compatibility is crucial, so if it's required for
the older versions, I agree not to remove that.

Regards,

-- 
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center