LiveUSB/LiveHDD Shutdown Error

[AS]

seems it can be done with aufs .....  but am failing to find anything about repeatable success with union fs 

http://aufs.sourceforge.net/

Aufs author/maintainer:
"Note: it becomes clear that "Aufs was rejected. Let's give it up." According to Christoph Hellwig, linux rejects all union-type filesystems but UnionMount."
 

[Was_Just19]

seems it can be done with aufs .....  but am failing to find anything about repeatable success with union fs 

http://aufs.sourceforge.net/

Aufs author/maintainer:
"Note: it becomes clear that "Aufs was rejected. Let's give it up." According to Christoph Hellwig, linux rejects all union-type filesystems but UnionMount."
 

  that seems to end that avenue 

[smcs_steve]

http://www.murga-linux.com/puppy/viewtopic.php?p=423346#423346
(May 2010)
scroll down to this section>

METHOD 2 : ROOTFS COMPLETE-THRU-THE-LAYERS UNMOUNT

Hope this helps,
>Steve

PS.
http://wiki.debian.org/AdvancedStartupShutdownWithMultilayeredBlockDevices
seems to suggest remounting / system file structure as read-only.....   can this be achieved with with the LiveUSB partition just before halt?

Code: [Select]

mount -n -o remount /??
  

[smcs_steve]

Memo to myself...

I think I'll ditch the whole persistence idea till a fix is found for this shutdown error (& the @/2 thing).
For the mean time I'll just make do with isohybrid of my remaster to the front end of the stick and fat32 partition for the rest!
(same as Mandriva does)
>Steve

[Was_Just19]

On flash stick, the unclean FS does not seem to affect the operation that I have noticed ......  but I may just not have noticed 

It has though called a halt to any idea of 'frugal'/poormans/liveHDD  installs with persistence.

[AS]

On flash stick, the unclean FS does not seem to affect the operation that I have noticed ......  but I may just not have noticed 

It has though called a halt to any idea of 'frugal'/poormans/liveHDD  installs with persistence.

Aren't problems made to be solved ? It will take a bit longer ... but it will be solved! 

AS

[Was_Just19]

On flash stick, the unclean FS does not seem to affect the operation that I have noticed ......  but I may just not have noticed 

It has though called a halt to any idea of 'frugal'/poormans/liveHDD  installs with persistence.

Aren't problems made to be solved ? It will take a bit longer ... but it will be solved! 

AS

Oh yes indeed 

All it means - until solved - is that persistence should not be used on liveHDD 

[smcs_steve]

On flash stick, the unclean FS does not seem to affect the operation that I have noticed ......  but I may just not have noticed  

It has though called a halt to any idea of 'frugal'/poormans/liveHDD  installs with persistence.

Not so in my tests -- eventually corruption creeps into data and config files in the changes area and makes the stick unbootable.  To remidy this I have found the changes folder needs to be deleted (or renamed to sideline the corruption).  Sometimes the changes folder cannot be deleted and the stick needs to be fsck in another linux system first, then it can be deleted.  This has been the case from the very beginning of PCLOS LiveUSB stick creation project -- in my experience.  Sad but true....
>Steve

[AS]

googling around searching for info (and status) about unionfs, aufs, unionmount, encountered a hyperlink: "unionfs in examples", from here: http://www.fsl.cs.sunysb.edu/project-unionfs.html and clicking on it:

"last, but not the least, AuFS is really stable in real world situations. I used unionfs in the past, but my second name for it was 'NULL POINTER DEREFERENCE'. I can see those errors still happening in latest unionfs 2.5.x as well, last one I've found in the mailing list was from 27th of April 2011 ..."
http://www.unionfs.org/

aufs info: http://aufs.sourceforge.net/aufs.html

a recent thread about unionmount status ... https://lkml.org/lkml/2011/4/12/510

After several days reading up and down on the net, my idea is that right now the best choice is aufs, of course may be I'm completely wrong about.

Apparently mylivecd script and pclos kernels are already providing aufs support, I'm going todo some testing around aufs "mylivecd --ufs aufs" in the next days...

AS

[Was_Just19]

Thanks for the links 

I do not know what the status of aufs is in the present kernels .......  IIRC some time ago it was troublesome .....  I hope I am not confusing it with something else 

[AS]

I'm going todo some testing around aufs "mylivecd --ufs aufs" in the next days...

So far, aufs seems to me perfectly equivalent to unionfs, successfully built a livecd ISO, successfully booted without problems. Things are a little different for liveHDD with or without persistence, where the main problem is that actually the rw partition is not unmounted at shutdown, leading to dirty filesystems either using unionfs or aufs. Therefore it's time to fix this umount issue, once for all. Tests go on, stay tuned ...

AS

[Was_Just19]

I'm going todo some testing around aufs "mylivecd --ufs aufs" in the next days...

So far, aufs seems to me perfectly equivalent to unionfs, successfully built a livecd ISO, successfully booted without problems. Things are a little different for liveHDD with or without persistence, where the main problem is that actually the rw partition is not unmounted at shutdown, leading to dirty filesystems either using unionfs or aufs. Therefore it's time to fix this umount issue, once for all. Tests go on, stay tuned ...

AS

That is good news indeed.
Have you had the opportunity to try LiveUSB with persistence using aufs?
If you do not have that facility, let me know and I can try it here from your instructions.

regards.

[AS]

Have you had the opportunity to try LiveUSB with persistence using aufs?
If you do not have that facility, let me know and I can try it here from your instructions.

No, liveUSB is the only I have not tested, but expect results similar to liveHDD: dirty partition at reboot; this issue actually can't allow to understand if filesystem inconsistency depends from lack of umount or from aufs/unionfs issues.
- kernel 2.6.38.x
- destination partiotion ext2/ext3 or ext4.
- mylivecd --ufs aufs --gzip test-aufs.iso

AS

[Was_Just19]

Yes the unclean unmount of the partition when using LiveUSB has been one of those things that has worried me for quite a while.

So using aufs will not change anything at present, but *should* allow changes to be made, that are not presently possible with unionfs, to eliminate the unclean unmounts of both partition and changes (if persistence is used).

Do I understand correctly?

I will later today (maybe tonight) look to creating a new ISO here using aufs and then try it out on LiveUSB.

regards.

[AS]

Yes the unclean unmount of the partition when using LiveUSB has been one of those things that has worried me for quite a while.

So using aufs will not change anything at present, but *should* allow changes to be made, that are not presently possible with unionfs, to eliminate the unclean unmounts of both partition and changes (if persistence is used).

Do I understand correctly?

I will later today (maybe tonight) look to creating a new ISO here using aufs and then try it out on LiveUSB.

regards.

Yes, you understood correctly.

What I'm trying to do right now, is to modify a little the /etc/init.d/halt script, to be able to log what is going on. Already seen that the script halt.local
is effectively executed only from a liveCD session, and is never executed from liveHDD or liveUSB installations.

AS