PostgreSQL and the OCFS2 filesystem - Mailing list pgsql-general

From Ploski, Karen L
Subject PostgreSQL and the OCFS2 filesystem
Date
Msg-id 19D0D50E9B1D0A40A9F0323DBFA04ACC0334C3B6@USRV-EXCH4.na.uis.unisys.com
Whole thread Raw
Responses Re: PostgreSQL and the OCFS2 filesystem  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general

The change log for Kernel Stable Build 2.6.16 indicates that OCFS2 has been integrated into the kernel.

(See http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.16  and http://www.linux-watch.com/news/NS3447236466.html).

I've searched the PostgreSQL mail archives for previous discussions of OCFS and OCFS2.  It appears the topic has come up before in various mail lists, including the PostgreSQL General and Administration mail lists. (For example: http://archives.postgresql.org/pgsql-admin/2004-04/msg00101.php).

At that time, there was concern about the OCFS/OCFS2 license (GPL) versus the PostgreSQL license (BSD) and the amount of work required to evaluate PostgreSQL's performance with this file system. 

Some of the discussions took place in the context of raw devices versus filesystems (http://archives.postgresql.org/pgsql-admin/2004-03/msg00376.php).

The discussion in at least two different threads suggested that shared memory was the real issue for PostgreSQL, rather than the filesystem (http://archives.postgresql.org/pgsql-general/2005-08/msg01181.php  and http://archives.postgresql.org/pgsql-general/2006-01/msg01244.php).

Now that OCFS2 is part of the kernel, the licensing issue would seem to have been put to rest.  Does the fact that OCFS2 is now part of the kernel help reduce the size of the work effort required to have a look at the possibility of PostgreSQL using OCFS2?  Is shared memory still the real issue? Any thoughts?

Thanks in advance,

Karen

THIS COMMUNICATION MAY CONTAIN CONFIDENTIAL AND/OR OTHERWISE PROPRIETARY MATERIAL and is thus for use only by the intended recipient. If you received this in error, please contact the sender and delete the e-mail and its attachments from all computers.

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Vacuuming of indexes on tables.
Next
From: elein
Date:
Subject: Partial indexes