[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180627123205.lckj7w3tcevqetcz@localhost.localdomain>
Date: Wed, 27 Jun 2018 14:32:05 +0200
From: Lukas Czerner <lczerner@...hat.com>
To: "Theodore Y. Ts'o" <tytso@....edu>
Cc: linux-ext4@...r.kernel.org, adilger@...ger.ca
Subject: Re: test f_extent_htree fails on 1.44.3-rc1
On Tue, Jun 26, 2018 at 06:41:07PM +0200, Lukas Czerner wrote:
> On Tue, Jun 26, 2018 at 11:52:00AM -0400, Theodore Y. Ts'o wrote:
> > On Tue, Jun 26, 2018 at 02:58:18PM +0200, Lukas Czerner wrote:
> > >
> > > Presumably because there is not enough blocks to allocate to create
> > > recreate the directory structure.
> > >
> > > This is the error I get from the given mke2fs
> > >
> > > Copying files into the device: __populate_fs: Could not allocate block in ext2 filesystem while writing file "yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy.202"
> > > mke2fs: Could not allocate block in ext2 filesystem while populating file system
> > >
> > > I am not really sure if the size of the file system plays any role in
> > > the test so can you take a look ?
> >
> > Thanks for reporting the bug!
> >
> > I suspect I know what's going on. The size of the file system is
> > calculated to be just large enough for files created in
> > /tmp/f_extent_htree.tmp/subdir, which get copied into the file system
> > using mke2fs -d /tmp/f_extent_htree.tmp/subdir.
> >
> > I'm guessing you have SELinux enabled, or something else which is
> > automatically creating extended attributes (e.g., IMA, etc.)?
>
> Right, I have selinux enabled and there are selinux labels created.
>
> >
> > Mke2fs -d will copy the xattrs into file system to be created, and
> > this is taking up more space than we had allowed for.
> >
> > We probably need to have a way to disable the xattr copy operation so
> > we get the same results on all systems. We probably also need a new
> > test which tests all of the new functionality in mke2fs -d, but that's
> > a different issue.
>
> Agreed.
Hi Ted,
just letting you know that indeed, when disabling set_inode_xattr() in
__populate_fs() the test runs ok. If you don't already have one I'll
work on a patch to add this option.
-Lukas
>
> Thanks!
> -Lukas
> >
> > - Ted
Powered by blists - more mailing lists