From mboxrd@z Thu Jan 1 00:00:00 1970 From: Theodore Ts'o Subject: Re: Max blocks per group? Date: Thu, 28 Jun 2012 15:23:03 -0400 Message-ID: <20120628192303.GA9054@thunk.org> References: <4FEC754F.5000907@ubuntu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-ext4@vger.kernel.org To: Phillip Susi Return-path: Received: from li9-11.members.linode.com ([67.18.176.11]:53346 "EHLO imap.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750769Ab2F1TXG (ORCPT ); Thu, 28 Jun 2012 15:23:06 -0400 Content-Disposition: inline In-Reply-To: <4FEC754F.5000907@ubuntu.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Thu, Jun 28, 2012 at 11:16:31AM -0400, Phillip Susi wrote: > > I notice that mke2fs refuses to allow more than 8 * block size blocks > per group. Why is this? Is it not permissible to have more than one > block for the block allocation bitmap? No, it's not permissible; it's been a fundamental design constraint in ext2/3/4 since the very beginning. You can, however, get the moral equivalent by using flex_bg which moves the metadata blocks to the beginning of an aggregated of (for example) 16 block groups of a flex_bg. The block allocation bitmaps are allocated contiguously, so between that and locating all of the metadata blocks of the bg's at the beginning of a flex_bg, that should get most of what you'd want with a larger number of blocks in a block group. And BTW, this is already the default in ext4 (using 16 block groups in each flex_bg). - Ted