Date: Thu, 23 May 2002 14:41:54 -0700 From: Sean Chittenden <sean@chittenden.org> To: ijliao@FreeBSD.org Cc: freebsd-ports@FreeBSD.org, freebsd-standards@freebsd.org Subject: Re: ports/36954: PostgreSQL daylight savings fix... Message-ID: <20020523144154.J68814@ninja1.internal> In-Reply-To: <200205230229.g4N2TqA84236@freefall.freebsd.org>; from "ijliao@FreeBSD.org" on Wed, May 22, 2002 at = 07:29:52PM References: <200205230229.g4N2TqA84236@freefall.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> Synopsis: PostgreSQL daylight savings fix... > > State-Changed-From-To: open->feedback > State-Changed-By: ijliao > State-Changed-When: Wed May 22 19:28:28 PDT 2002 > State-Changed-Why: > maintainer said that he cant reproduce this bug > could you please try again ? > or does this bug depend on any env vars ? > > http://www.freebsd.org/cgi/query-pr.cgi?pr=36954 Ugh. This topic is a PITA. 1) FreeBSD is technically correct in that 2AM doesn't exist at a daylight savings switch (according to Naval time sources/reference/documentation). 2) FreeBSD is the _only_ operating system that returns an error for this time value. For consistency sake, should postgresql fudge over the error or should FreeBSD's mktime() behave like the rest of the *NIX's, including Solaris, AIX, and our favorite bastard child, Linux? ::shrug:: This discussion never seems to end. So I patched PostgreSQL and had a long discussion with them about this (see bug PG #630). This discussion comes up every spring, but looses its importance every year after the poor flag carrier gets thumped into submission. To quote from the discussion: "Personally I'd like to see FreeBSD do away with this strange behavior. It cause my grief because certain things *MUST* be done at 0200 every day in our system, I was forced to do them manually recently, shifting several hours of work into daytime which had to be paused and bulked into the next days work. I realize that this is getting off track but it just points out that the FreeBSD behavior is IMHO WRONG. It causes applications to fail in an unexpected and odd way. "I'm not objecting to pg patching for it (no choice at the moment) but I hope the pg team 'officially' puts a little pressure on the BSD folk to make this behave as expected. "I don't have any compliance docs at the moment, but this strikes me as somewhat out of spec personally." And so I'm stuck as to what to do. I'm not the person in the quote, but have the same problem (really a cron issue, but that's neither here nor there). Every hour, on the hour, something happens and it involves inserting data and every year, it fails for only PG databases on FreeBSD. The problem is quite reproducible, I've done it on numerous occasions and if anyone has any problems, please let me know. I'd like to see this patch included if possible and hopefully something worked out on freebsd-standards (though I've kinda dropped the ball on keeping that discussion alive because I got tired of it). Anyone else want to help carry the flag? <:~) -sc -- Sean Chittenden To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-standards" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020523144154.J68814>