From mboxrd@z Thu Jan 1 00:00:00 1970 From: PFC Subject: fsck --rebuild-tree just because my battery ran out... Date: Sun, 30 Oct 2005 00:21:03 +0200 Message-ID: References: <20051028223547.GF4180@stusta.de> <20051029092250.GG4180@stusta.de> <43639EA0.5080300@namesys.com> Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Return-path: list-help: list-unsubscribe: list-post: Errors-To: flx@namesys.com In-Reply-To: <43639EA0.5080300@namesys.com> List-Id: Content-Type: text/plain; format="flowed"; delsp="yes"; charset="us-ascii" To: Hans Reiser , Alexander Zarochentcev Cc: reiserfs-list@namesys.com I do not know if a laptop battery running out behaves the same as a power cord yanking, but this fact may be interesting : As I was just reading some document, with no other activity, the harddrive had spinned down since a while, and was NOT running (as in NOT spinning) when the battery ran out. So it seems, the HD had spinned down, while the FS was in an inconsistent state. Strange. This happened several times, out of which, once I had to fsck, and the other to --rebuild-tree. linux-2.6.11-cko1-swsusp2 reiser4 gentoo Centrino Pentium M laptop cat /proc/sys/vm/laptop_mode 0 /dev/hda: multcount = 16 (on) IO_support = 0 (default 16-bit) unmaskirq = 0 (off) using_dma = 1 (on) keepsettings = 0 (off) readonly = 0 (off) readahead = 256 (on) geometry = 65535/16/63, sectors = 60011642880, start = 0 I assume the disk flushes its write cache before going to standby (or else the manufacturers are crazy)... I hope this can help... >> Remember my message where I said reiser4 was preventing my laptop >> from shutting down itself ? >> >> Well it turned out that, in fact, it was a stupid shfs-mount >> process which would refuse to die. Adding a "killall ssh" in the >> local.stop script is a bit violent, but perfectly functional way, to >> bring it back to reason. Everything now works. >> >> But still, I had to fsck --rebuild-tree just because my battery >> ran out... >> >> >> > the battery running out should not have caused you to need to do that. > > zam, can you do some testing of how well we handle power cord yanking > while under load? There are some guys at stanford who have some test > programs for this sort of thing, I wish I could remember their name at > the moment....http://www.materiel.net/details_VA7000BWA.html > > Hans >