From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Wed, 7 Jun 2000 01:03:13 +0100 (IST) From: Paul Jakma Subject: Re: [linux-lvm] LVM 0.8final for 2.2.15/2.2.16? In-Reply-To: <20000607004618.A13524@gruyere.muc.suse.de> Message-ID: MIME-Version: 1.0 Sender: owner-linux-lvm Errors-To: owner-linux-lvm List-Id: Content-Type: TEXT/PLAIN; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Andi Kleen Cc: Linux LVM list On Wed, 7 Jun 2000, Andi Kleen wrote: > You cannot safely access a unclean file system without replaying the log > (when it is journaled) or running fsck (when it is not journaled). The > file system meta data may not be consistent: but why would the snapshot capture the fs in an inconsistent state? > files may contain bogus > blocks, directories may point to nowhere, inodes may be outdated etc. only if the fs is inconsistent.. but why should the fs be inconsistent when making the snapshot? aiui the snapshot works within the lvm block layer. the fs works on the block layer. for the fs to be inconsistent there must be a disparity between the fs and the block layer at snapshot time, right? but surely the snapshot would do some kind of atomic sync at the point of snapshot creation? (what am missing?) > In short, without write access you cannot safely read it, except > when it was cleanly unmounted > would mount -o remount,sync ; lvcreate -s ... ; mount -o remount,async work? > > -Andi > > -- Paul Jakma paul@clubi.ie PGP5 key: http://www.clubi.ie/jakma/publickey.txt ------------------------------------------- Fortune: Democracy is the recurrent suspicion that more than half of the people are right more than half of the time. -- E. B. White