Thread: Specifying xlog directory during initdb is failing
Hi, I was testing -X option, and I cannot make it work. This is PostgreSQL 8.3.1 on Fedora 9: ========================================================== -bash-3.2$ initdb -d -X xlog/ -D data/ Running in debug mode. VERSION=8.3.1 PGDATA=data share_path=/usr/share/pgsql PGPATH=/usr/bin POSTGRES_SUPERUSERNAME=postgres POSTGRES_BKI=/usr/share/pgsql/postgres.bki POSTGRES_DESCR=/usr/share/pgsql/postgres.description POSTGRES_SHDESCR=/usr/share/pgsql/postgres.shdescription POSTGRESQL_CONF_SAMPLE=/usr/share/pgsql/postgresql.conf.sample PG_HBA_SAMPLE=/usr/share/pgsql/pg_hba.conf.sample PG_IDENT_SAMPLE=/usr/share/pgsql/pg_ident.conf.sample The files belonging to this database system will be owned by user "postgres". This user must also own the server process. The database cluster will be initialized with locale en_US.UTF-8. The default database encoding has accordingly been set to UTF8. The default text search configuration will be set to "english". creating directory data ... ok creating directory xlog/ ... ok creating subdirectories ... initdb: could not create directory "data/pg_xlog": File exists initdb: removing data directory "data" initdb: removing transaction log directory "xlog" ========================================================== Am I missing something here? If not, this is a bug I believe. Regards, -- Devrim GÜNDÜZ , RHCE PostgreSQL Replication, Consulting, Custom Development, 24x7 support Managed Services, Shared and Dedicated Hosting Co-Authors: plPHP, ODBCng - http://www.commandprompt.com/
Devrim GÜNDÜZ <devrim@CommandPrompt.com> writes:
> I was testing -X option, and I cannot make it work. This is PostgreSQL
> 8.3.1 on Fedora 9:
> -bash-3.2$ initdb -d -X xlog/ -D data/
The root of the problem is that you can't use a relative path for -X:
it's not clear whether the path is relative to $PGDATA or the working
directory when you invoked initdb, and different bits of initdb
effectively make each of those assumptions. I agree that the failure
mode is pretty opaque :-(. I think the simplest solution is to reject
non-absolute path for -X; is there any real use-case for allowing it?
regards, tom lane
On Sun, Jun 01, 2008 at 10:27:25PM -0400, Tom Lane wrote: > mode is pretty opaque :-(. I think the simplest solution is to reject > non-absolute path for -X; is there any real use-case for allowing it? No. A -- Andrew Sullivan ajs@commandprompt.com +1 503 667 4564 x104 http://www.commandprompt.com/
Devrim GÜNDÜZ <devrim@CommandPrompt.com> writes:
> On Sun, 2008-06-01 at 22:27 -0400, Tom Lane wrote:
>> I think the simplest solution is to reject non-absolute path for -X;
>> is there any real use-case for allowing it?
> I don't think so.
> Thanks for the patch -- could you please update docs (and man page) as
> well? There are no words regarding using absolute path there, AFAICS.
Don't you think the error message will be sufficient?
regards, tom lane
Hi, On Mon, 2008-06-02 at 01:16 -0400, Tom Lane wrote: > > Thanks for the patch -- could you please update docs (and man page) > as > > well? There are no words regarding using absolute path there, > AFAICS. > > Don't you think the error message will be sufficient? I would not even try using non-absolute directory if it was specified in the docs ;) Regards, -- Devrim GÜNDÜZ , RHCE PostgreSQL Replication, Consulting, Custom Development, 24x7 support Managed Services, Shared and Dedicated Hosting Co-Authors: plPHP, ODBCng - http://www.commandprompt.com/
Hi, On Sun, 2008-06-01 at 22:27 -0400, Tom Lane wrote: > I think the simplest solution is to reject non-absolute path for -X; > is there any real use-case for allowing it? I don't think so. Thanks for the patch -- could you please update docs (and man page) as well? There are no words regarding using absolute path there, AFAICS. Regards, -- Devrim GÜNDÜZ , RHCE PostgreSQL Replication, Consulting, Custom Development, 24x7 support Managed Services, Shared and Dedicated Hosting Co-Authors: plPHP, ODBCng - http://www.commandprompt.com/
Hi, On Mon, 2008-06-02 at 08:21 +0300, Devrim GÜNDÜZ wrote: > > Don't you think the error message will be sufficient? > > I would not even try using non-absolute directory if it was specified > in the docs ;) BTW: We do allow non-absolute directory names for $PGDATA, but don't allow for xlogs -- it does not look consistent. That's why I'm asking for a doc update. Regards, -- Devrim GÜNDÜZ , RHCE PostgreSQL Replication, Consulting, Custom Development, 24x7 support Managed Services, Shared and Dedicated Hosting Co-Authors: plPHP, ODBCng - http://www.commandprompt.com/