Re: pg_basebackup option for handling symlinks

From: Magnus Hagander <magnus(at)hagander(dot)net>
To: Peter Eisentraut <peter_e(at)gmx(dot)net>
Cc: pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: pg_basebackup option for handling symlinks
Date: 2012-01-20 14:34:11
Message-ID: CABUevEzzFpwUqgptf0Hu_enZwtsNQb6bbbRGQ0A-f8HYoUUeTA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Jan 16, 2012 at 19:52, Peter Eisentraut <peter_e(at)gmx(dot)net> wrote:
> On sön, 2012-01-08 at 22:22 +0100, Magnus Hagander wrote:
>> On Sun, Jan 8, 2012 at 21:53, Peter Eisentraut <peter_e(at)gmx(dot)net> wrote:
>> > I've recently had a possible need for telling pg_basebackup how to
>> > handle symlinks in the remote data directory, instead of ignoring them,
>> > which is what currently happens.  Possible options were recreating the
>> > symlink locally (pointing to a file on the local system) or copying the
>> > file the symlink points to.  This is mainly useful in scenarios where
>> > configuration files are symlinked from the data directory.  Has anyone
>> > else had the need for this?  Is it worth pursuing?
>>
>> Yes.
>>
>> I came up to the same issue though - in one case it would've been best
>> to copy the link, in the other case it would've been best to copy the
>> contents of the file :S Couldn't decide which was most important...
>> Maybe a switch would be needed?
>
> Yes.  Do we need to preserve the third behavior of ignoring symlinks?

I don't think we do.

> tar has an -h option that causes symlinks to be followed.  The default
> there is to archive the symlink itself.

Seems like a reasonable pattern to follow (though I think using -h is
a really bad idea, but the pattern of by default archiving the symlink
itself)

--
 Magnus Hagander
 Me: http://www.hagander.net/
 Work: http://www.redpill-linpro.com/

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Simon Riggs 2012-01-20 14:34:31 Re: Online base backup from the hot-standby
Previous Message Heikki Linnakangas 2012-01-20 14:29:28 Re: our buffer replacement strategy is kind of lame