Обсуждение: BUG #6389: Fsync failed while creating DB

The following bug has been logged on the website:

Bug reference:      6389
Logged by:          Anjali Arora
Email address:      anjali_524@yahoo.co.in
PostgreSQL version: 9.0.4
Operating system:   CentOS
Description:=20=20=20=20=20=20=20=20

It is reference to bug#6372..

Hi Tom,

=20

Thanks for the solution. CIFS worked with fsync flag by ingnoring EINVAL in
copydir.c.=20

=20

I tested fsync with 8.2.2 version of PostgreSQL, it worked fine without
EINVAL patch. I wanted to know is something changed in version 9.0.4 of
postgreSQL.

=20

As fsync flag was not working with PostgreSQL version 9.0.4 without applying
the patch.

=20

Regards,

Anjali

On Tue, Jan 10, 2012 at 1:02 AM,  <anjali_524@yahoo.co.in> wrote:
> Thanks for the solution. CIFS worked with fsync flag by ingnoring EINVAL in
> copydir.c.
>
> I tested fsync with 8.2.2 version of PostgreSQL, it worked fine without
> EINVAL patch. I wanted to know is something changed in version 9.0.4 of
> postgreSQL.
>
> As fsync flag was not working with PostgreSQL version 9.0.4 without applying
> the patch.

Sure, lots of stuff has changed since REL8_2_2 - that release is
almost 5 years old.  I'm not sure it's really important to drill down
and find out exactly which change introduced this problem, but you
could start by trying 8.2.23, 8.3.x, 8.4.x, etc. if you want to try to
figure it out.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Robert Haas <robertmhaas@gmail.com> writes:
> On Tue, Jan 10, 2012 at 1:02 AM,  <anjali_524@yahoo.co.in> wrote:
>> Thanks for the solution. CIFS worked with fsync flag by ingnoring EINVAL in
>> copydir.c.
>>
>> I tested fsync with 8.2.2 version of PostgreSQL, it worked fine without
>> EINVAL patch. I wanted to know is something changed in version 9.0.4 of
>> postgreSQL.
>>
>> As fsync flag was not working with PostgreSQL version 9.0.4 without applying
>> the patch.

> Sure, lots of stuff has changed since REL8_2_2 - that release is
> almost 5 years old.  I'm not sure it's really important to drill down
> and find out exactly which change introduced this problem, but you
> could start by trying 8.2.23, 8.3.x, 8.4.x, etc. if you want to try to
> figure it out.

The reason 8.2 appears to "work" here is that it doesn't try to fsync
the output at all during CREATE DATABASE.  It is not a bug that we added
fsync logic.  The only thing that's in question is whether we should
allow certain errno values to pass without comment.

            regards, tom lane