[pgpool-general: 2078] Re: unexpected node degeneration

[Tatsuo Ishii]

> Hi,
> 
> I have found that I can reliably cause a backend to degenerate by
> creating a database using pgadmin3.  Here is the pgpool-II logging:
> 
> 2013-08-26 10:23:50 ERROR: pid 30038: read_kind_from_backend: 0 th
> kind D does not match with master or majority connection kind C
> 2013-08-26 10:23:50 ERROR: pid 30038: kind mismatch among
> backends. Possible last query was: "SET DateStyle=ISO;
> SET client_min_messages=notice;
> SET bytea_output=escape;
> SELECT oid, pg_encoding_to_char(encoding) AS encoding, datlastsysoid
>   FROM pg_database WHERE oid = 193693" kind details are: 0[D] 1[C] 2[C]
> 2013-08-26 10:23:50 LOG: pid 30038: degenerate_backend_set: 0 fail
> over request from pid 30038
> 
> This is with pgpool-II 3.3.0.

Yes, it's a known problem. Pgpool-II cannot correctly handle multi
statement. It treats the multi statement as a write query which causes
that kind of problem.
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese: http://www.sraoss.co.jp