From: Eric Sandeen Subject: Re: [PATCH] e2image: Print a warning if running over a mounted filesystem Date: Thu, 26 Sep 2013 19:39:30 -0500 Message-ID: <5244D3C2.4090606@redhat.com> References: <1380229204-32526-1-git-send-email-cmaiolino@redhat.com> <20130926235658.GD6011@thunk.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Carlos Maiolino , linux-ext4@vger.kernel.org To: "Theodore Ts'o" Return-path: Received: from mx1.redhat.com ([209.132.183.28]:3822 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752808Ab3I0Ajd (ORCPT ); Thu, 26 Sep 2013 20:39:33 -0400 In-Reply-To: <20130926235658.GD6011@thunk.org> Sender: linux-ext4-owner@vger.kernel.org List-ID: On 9/26/13 6:56 PM, Theodore Ts'o wrote: > On Thu, Sep 26, 2013 at 06:00:04PM -0300, Carlos Maiolino wrote: >> Several users use to run e2image over a mounted filesystem, providing >> inconsistent, useless e2images. >> This patch adds a warning in such cases, notifying the user and also adds a >> force option making e2image able to run over Read-only filesystems. > > It should be perfectly safe to run e2image on a read-only mounted file > system option, so it's not obvious to me why the force option would be > needed in that case. right now Carlos' test isn't checking for readonly; just mounted, right? But I think it should check for ro, and allow it by default in that case, I agree. > Also, if we are saving a "normal" (not a raw or qcow) e2image file, we > are only backing up the statically located metadata blocks (i.e., > superblock, block group descriptors, inode table, and allocation > bitmaps). If we do this on a mounted file system, the e2image file is > less useful, but I'm not sure I'd call it completely useless. If the > goal is to backup critical metadata, it will do that just fine. So > maybe it's worthy of a warning, but I'm not sure it should require a > force option. I asked Carlos to do this after getting the 2nd customer filesystem image in a week which was useless for triage due to having been run on a live, mounted fs... I fear that a warning would be ignored, but *shrug* - at least something so we have some hope of getting something useful out the other end of e2image -r or -Q... TBH I've never used a "normal" image; if you want to allow it to run on an RW filesystem that won't bother me at all. ;) > If the user is trying to capture a raw or qcow image file, I agree > that requiring that the file systme either be mounted read-only, not > mounted at all, or that a force option be specified, makes sense. cool. :) -Eric > - Ted > -- > To unsubscribe from this list: send the line "unsubscribe linux-ext4" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html >