[pgpool-general: 2065] Re: trouble with recovery of a downed node

[Sean Hogan]

On 13-08-21 08:23 PM, Tatsuo Ishii wrote:
>> Ah, I believe I understand the problem now: the sequence values are 
>> different (off by one) between the original and recovered databases. 
>> The sample pgpool_sample_pitr script has: SELECT setval(oid, 
>> nextval(oid)) FROM pg_class WHERE relkind = 'S' When I change that to 
>> SELECT setval(oid, 1*currval(oid)) FROM pg_class WHERE relkind = 'S' 
>> I get identical databases and the problem has not reoccurred! (The 
>> "1*" is to defeat optimization; not sure if it is needed.) Thanks for 
>> your excellent advice. 
> Thanks for the report but your changes seems a little bit strange to
> me. According to the PostgreSQL manual, calling currval() without
> prior calling nextval() will raise an error. Don't you get any error?
> --
> Tatsuo Ishii
> SRA OSS, Inc. Japan
> English: http://www.sraoss.co.jp/index_en.php
> Japanese: http://www.sraoss.co.jp

That's interesting, I was unaware of that restriction on currval. Indeed 
if I run the script manually I do get an error.  Now I'm quite confused, 
because the overall effect is what I need.  Based on the manual page at 
http://www.postgresql.org/docs/9.2/static/functions-sequence.html if 
appears the following will work:

SELECT setval(oid, nextval(oid), false) FROM pg_class WHERE relkind = 'S'

and my initial experiments suggest it does.

Sean
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pgpool.net/pipermail/pgpool-general/attachments/20130822/4a0d07fb/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sean.vcf
Type: text/x-vcard
Size: 275 bytes
Desc: not available
URL: <http://www.pgpool.net/pipermail/pgpool-general/attachments/20130822/4a0d07fb/attachment.vcf>