From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tomasz Chmielewski Subject: Re: [linux-lvm] Q: Online resizing ext3 FS Date: Wed, 12 Sep 2007 16:56:50 +0200 Message-ID: <46E7FE32.7010404@wpkg.org> References: <20070912163621.1f951876@slopi> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20070912163621.1f951876@slopi> Sender: linux-raid-owner@vger.kernel.org To: LVM general discussion and development Cc: linux-raid@vger.kernel.org List-Id: linux-raid.ids Chris Osicki schrieb: > Hi > > I apologize in advance for asking a question not really appropriate > for this mailing list, but I couldn't find a better place with lots of > people managing lots of disk space. > > The question: > Has anyone of you been using ext2online to resize (large) ext3 filesystems? > I have to do it going from 500GB to 1TB on a productive system I was > wondering if you have some horror/success stories. > I'm using RHEL4/U4 (kernel 2.6.9) on this system. Yes, I tried to online resize a similar filesystem (600 MB to 1.2 TB) and it didn't work. At some point, resize2fs would just exit with errors. I tried to do it several times before I figured out what's missing; sometimes, I interrupted the process with ctrl+c. No data loss occurred. To do an online ext3 resize, the filesystem needs a "resize_inode" feature. You can check the features with dumpe2fs: # dumpe2fs -h /dev/sda1 (...) Filesystem features: has_journal resize_inode dir_index filetype needs_recovery sparse_super large_file (...) This flag is added by default only in the recent versions of e2progs (1.39 and later AFAIR); before, it had to be specified manually. So with RHEL4, you may be out of luck. In the end, I had to to an offline resize. I had this volume mirrored on another machine, so I didn't worry that much though. Also, to resize a filesystem of that size you would need plenty of RAM (if you have about 1 GB RAM free, it should be just enough; otherwise, your machine will be swapping, and the process will take longer). Before, I tried to resize it on a machine with 256 MB and several snapshots; resize2fs was killed because of OOM, and still, no data loss. If you have that an old kernel, take care if you're using snapshots; I believe they are stable only as of 2.6.22 (before 2.6.22 snapshots needed a lot of RAM; before 2.6.18 there were problems with snapshots removing etc.). Would be good to add some of that info to LVM HOWTO. -- Tomasz Chmielewski http://wpkg.org