[dm-crypt] xfs and fstab. To HeinzDiehl, first of all. Ende
sigbj-st at operamail.com
Sat Jan 9 18:15:10 CET 2010
OKEY! Will say this is finished, but for personal interest just go ahead and read some last comments on your last answer. Thank you my friend!
/* Was meistern und vermögen wir ohne die Hilfe Karlheinz Diehls. */
> ----- Original Message -----
> From: "Heinz Diehl" <htd at fancy-poultry.org>
> To: dm-crypt at saout.de
> Subject: Re: [dm-crypt] xfs and fstab. To HeinzDiehl, first of all.
> Date: Sat, 9 Jan 2010 13:14:50 +0100
> On 09.01.2010, Si St wrote:
> > _dev_hdc6 /dev/hdc6 none luks
> This is the correct line for your partitions in crypttab, if you wish to
> be asked for a password in the booting phase to unlock the partition.
> If there's a message telling you "unknown option luks", I would consider
> it as a bug in the boot.crypto system unless proved otherwise and opened a
> bug report for it. Is there a possibility to upgrade?
Could be an upgrade somewhere, but SLED 10 is replaced from Novell with SLED 11 to-day, which does not work on my printers. They have followed the arch of openSuSE 11.1 which has the same problem for me.
> > /var/log/messages:
> > Jan 7 17:19:51 linux-wh8s kernel: XFS: bad magic number
> > Jan 7 17:19:51 linux-wh8s kernel: XFS: SB validate failed
> This indicates a major problem with the XFS superblock. XFS can't operate
> on this drive without having this fixed.
Only messages on the 7th of January had these errors. From 8th on the are absent. It seem to happen on the second nachfolgende partition with xfs. - openLuks mounted it manually.
> > /var/log/warn:
> > This came from the trying to open xfs on a filecontainer (note Dec 17):
> > Dec 17 15:22:54 linux-wh8s kernel: XFS: Filesystem dm-3 has duplicate UUID -
> > can't mount
> A process has already mounted the filesystem (or is attempting that right
> now) and has reserved the uuid. You'll have to find out what's the cause.
To avoid this the man mount gives an answer:
Ignore the filesystem uuid. This avoids errors for duplicate uuids."
Your reference page mentions similar things:
" - nouuid
Don't check for double mounted file systems using the file system uuid. This is useful to mount LVM snapshot volumes."
Solution: replace acl,user-xattr, with nouuid
> > The /var/log/messages notified that the 'acl' was unknown as to the XFS:
> > Jan 8 17:46:45 linux-wh8s kernel: XFS: unknown mount option [user_xattr].
> These are illegal XFS mount options and can be the cause that the underlying
> filesystem doesn't get mounted.
YES, you are right! acl,user_xattr, belongs only to ext2. My fault here.
As a starting point, Please read here:
> > Now, the xfs-partition was opened and there was error message
> > promptwise or in
> > the /var/log/messages:
> So I can consider your problem solved? That's fine!
> > Could I also have specified better the mkfs.xfs commandline too?
> > I just have done: mkfs.xfs /dev/hd..
> That's a matter of opinion and circumstances. Here's what I'm using,
> regardless of encryption:
> mkfs.xfs -f -l lazy-count=1,version=2 -i attr=2 -d agcount=6 /dev/xxx
> In fstab, I mount with
> Please keep in mind that the "nobarrier" option gives great speed and
> latency, but could cause severe data loss in case of a crash or power
> failure. My systems are backuped by an UPS. (And I can remember having
> read "barriers not supported" in your logs, this means the same).
These utterings are interesting. XFS is advanced as to configs and I had not known too much about it. The main reason for choice of XFS is that average file size in the mountdir for xfs is +4MB. Since xfs have advantages of large files (like jfs) I chose it.
On my machine I have 3 harddisks.
The intention is to have backup of my office on 2 archives - ark1 and ark2 that resides on 2 different harddisks and not only on 2 different partitions since painful experience is that the whole harddisk becomes inaccessible on many types of crashes. I tried first file-containers, then partitions. When I am ready I go for it on hda(/ark1/) and hdb(/ark2/) as backups. Not hdc9-10 that is only for test-purposes. Main system would be on hdc.
Here is an old archive on my home machine. Backup is done every day. Every month last backup for the month is saved as to compressed archives. An incremental archive is growing all the time containing all changes even if information is deleted by mistake. Thus recovery of deleted info is regained with tar -xOf arkiv.tar ./n-name.
sigbj at linux:~> sudo /usr/bin/tilkoble_xfs_krypt
skriv inn dekrypteringspassordet
Activating crypto devices using /etc/cryptotab ...
Please enter passphrase for /dev/hda5:
fsck 1.38 (30-Jun-2005)
/sbin/fsck.xfs: XFS file system.
sigbj at linux:~> cd /xfs_krypt/arkiv/
sigbj at linux:/xfs_krypt/arkiv> ls -ltr
sigbj users 0 2006-03-26 16:27 kon_prøve.tgz.gml
sigbj users 4096 2008-07-31 14:55 2007-A
sigbj users 4096 2008-07-31 14:55 2006-A
sigbj users 501 2008-07-31 15:36 LES
sigbj users 4096 2009-01-15 23:03 2008-A
sigbj users 3940808 2009-01-31 12:46 kon_Jan,2009.tgz.gml
sigbj users 3973520 2009-02-28 13:06 kon_Feb,2009.tgz.gml
sigbj users 4015591 2009-03-31 15:31 kon_Mär,2009.tgz.gml
sigbj users 4066579 2009-04-30 17:12 kon_Apr,2009.tgz.gml
sigbj users 4111358 2009-05-31 16:48 kon_Mai,2009.tgz.gml
sigbj users 4155993 2009-06-30 14:13 kon_Jun,2009.tgz.gml
sigbj users 4138583 2009-07-31 14:53 kon_Jul,2009.tgz.gml
sigbj users 4173209 2009-08-31 19:42 kon_Aug,2009.tgz.gml
sigbj users 4205100 2009-09-30 17:08 kon_Sep,2009.tgz.gml
sigbj users 4253602 2009-10-31 17:16 kon_Okt,2009.tgz.gml
sigbj users 4293112 2009-11-30 17:16 kon_Nov,2009.tgz.gml
sigbj users 24064000 2009-12-01 16:44 arkiv.tar
sigbj users 4303738 2009-12-01 16:45 kon_Dez,2009.tgz
sigbj at linux:/xfs_krypt/arkiv>
> dm-crypt mailing list
> dm-crypt at saout.de
Surf the Web in a faster, safer and easier way:
Download Opera 9 at http://www.opera.com
Powered by Outblaze
More information about the dm-crypt