On Tue, 6 May 2014, Theodore Ts'o wrote: > Date: Tue, 6 May 2014 11:21:29 -0400 > From: Theodore Ts'o <tytso@xxxxxxx> > To: Lukáš Czerner <lczerner@xxxxxxxxxx> > Cc: Darrick J. Wong <darrick.wong@xxxxxxxxxx>, linux-ext4@xxxxxxxxxxxxxxx > Subject: Re: [PATCH 37/37] ext5: define new subtype to add features and reduce > testing complexity > > On Tue, May 06, 2014 at 02:50:39PM +0200, Lukáš Czerner wrote: > > meta_bg - Just makes group descriptors to be spread across the file > > system. It has been around for some time and I am not sure > > why this is not a default already. It should also increase > > the limit of the file system size but I am not sure whether > > this is still true with flex_bg ? > > meta_bg signiicantly slows down mount operations (and in general, any > operation where we need to read in the block group descriptors --- > i.e., dumpe2fs, e2fsck, etc.) > > The strategy for meta_bg is that it's something that we enable as we > need it, as part of an online or off-line resize. That way, we keep > the block groups contiguous for as long as possible. Once the resize > inode has been exhausted (which _will_ happen when the file system > size grows beyond 16T), the resize operation will turn off the > resize_inode feature and then enable the meta_bg feature. > > And this is all working today, with the latest kernel and e2fsprogs; > so there's no reason to enable meta_bg as part of mke2fs operation, > and a good reason not to enable it by default, but to let resize2fs > turn it on when it makes sense to do so. > > - Ted Perfect, thanks for explanation. I was not very sure about meta_bg myself. Thanks! -Lukas
