Re: DROP DATABASE vs patch to not remove files right away

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Heikki Linnakangas <heikki(at)enterprisedb(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: DROP DATABASE vs patch to not remove files right away
Date: 2008-04-17 23:56:16
Message-ID: 1152.1208476576@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Heikki Linnakangas <heikki(at)enterprisedb(dot)com> writes:
> Patch attached that does the three changes we've talked about:'
> - make ForgetDatabaseFsyncRequests forget unlink requests as well
> - make rmtree() not fail on ENOENT
> - force checkpoint on in dropdb on all platforms

This looks fine as far as it goes, but I'm still thinking it's a bad
idea for rmtree() to fall over on the first failure. I propose that it
ought to keep trying to delete the rest of the target directory tree.
Otherwise, one permissions problem (for example) could lead to most
of a dropped database not getting freed up,

If there aren't objections, I'll make that happen after Heikki applies
his patch.

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2008-04-18 00:10:56 Re: Lessons from commit fest
Previous Message Gregory Stark 2008-04-17 23:55:30 Re: RFD: hexstring(n) data type