From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: WDC WD15EARS-00Z5B1: harddrive which does crazy unloading Date: Thu, 11 Mar 2010 14:53:52 +0900 Message-ID: <4B988570.4040005@kernel.org> References: <201002081933.10144.suse-linux@ml082.pinguin.uni.cc> <4B96013F.1050904@kernel.org> <4B965DE7.7000109@teksavvy.com> <4B9881C8.10506@kernel.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Return-path: Received: from hera.kernel.org ([140.211.167.34]:36373 "EHLO hera.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751631Ab0CKFx4 (ORCPT ); Thu, 11 Mar 2010 00:53:56 -0500 In-Reply-To: <4B9881C8.10506@kernel.org> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Mark Lord Cc: Al Bogner , linux-ide@vger.kernel.org On 03/11/2010 02:38 PM, Tejun Heo wrote: > Heh, yeah, that's pretty low. I hope they eventually learn to adjust > that dynamically considering running avg of intervals between writes. > It shouldn't be too hard. or just bump load cycle limit by x10 or x100. :-P storage-fixup can't be the permanent solution. The only reason I created it was because I thought it was a passing phase and the blacklist wouldn't keep growing. In both reports here, the avg load cycle frequency seems to be about 1 per 2 minutes. With 300k cycle rating (drives often happily go over it), it means 1.14 year of continuous operation, which depending on usage pattern might be considered acceptable. Maybe the right thing to do is to simply let it be and if it expires before within the warranty period, get it RMA'd. Hmmm... or, we can make a utility which monitors smart output, detects unload cycle and periodically poll load status and issue a bogus read whenever the disk unloads its heads so that we can storm the vendors with RMAs. :-) Coming back to the reality, I'll ask vendors about it and see whether we need to keep worrying about it or we can just let it be. Thanks. -- tejun