X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.144]) by svr1.postgresql.org (Postfix) with ESMTP id 6B92E5E3632 for ; Thu, 26 Aug 2004 17:01:23 -0300 (ADT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (av.hub.org [200.46.204.144]) (amavisd-new, port 10024) with ESMTP id 94354-10 for ; Thu, 26 Aug 2004 20:01:20 +0000 (GMT) Received: from wren.rentec.com (wren.rentec.com [65.213.84.9]) by svr1.postgresql.org (Postfix) with ESMTP id 16A075E3639 for ; Thu, 26 Aug 2004 17:01:17 -0300 (ADT) Received: from [172.26.132.145] (hoopoe.rentec.com [172.26.132.145]) by wren.rentec.com (8.12.9/8.12.1) with ESMTP id i7QK1HuY010349; Thu, 26 Aug 2004 16:01:17 -0400 (EDT) Message-ID: <412E418B.3070103@rentec.com> Date: Thu, 26 Aug 2004 16:01:15 -0400 From: Alan Stange Reply-To: stange@rentec.com Organization: Renaissance Technologies Corp. User-Agent: Mozilla Thunderbird 0.7.1 (X11/20040701) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Doug McNaught Cc: pgsql-performance@postgresql.org Subject: Re: Anyone familiar with Apple Xserve RAID References: <1093468177.12331.424.camel@camel> <200408251422.08366.josh@agliodbs.com> <08139423-F6EB-11D8-ACDB-000393DB553C@bahn.de> <200408251649.02932.josh@agliodbs.com> <412E34B8.6000708@speedfc.com> <87hdqpzndv.fsf@asmodeus.mcnaught.org> In-Reply-To: <87hdqpzndv.fsf@asmodeus.mcnaught.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at hub.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests= X-Spam-Level: X-Archive-Number: 200408/376 X-Sequence-Number: 8012 Doug McNaught wrote: >Kevin Barnard writes: > > > >> Actually you are both are right and wrong. The XRaid uses >> FibreChannel to communicate to the host machine(s). The Raid >> controller is a FibreChannel controller. After that there is a >> FibreChannel to UltraATA conversion for each drive, separate ATA bus >> for each drive. >> What I am curious about is if this setup gets around ATA fsync >> problems, where the drive reports the write before it is actually >> performed. >> >> > >Good point. > >(a) The FC<->ATA unit hopefully has a battery-backed cache, which > would make the whole thing more robust against power loss. >(b) Since Apple is the vendor for the drive units, they can buy ATA > drives that don't lie about cache flushes. Whether they do or not > is definitely a question. ;) > > FYI: http://developer.apple.com/technotes/tn/pdf/tn1040.pdf a tech note on write cache flushing. A bit dated now, but perhaps some other tech note from Apple has more recent information. -- Alan