Bug 22150

on mga6-64

packages installed cleanly
- btrfs-progs-4.14-2.mga6.x86_64
- lib64btrfs0-4.14-2.mga6.x86_64
- lib64zstd1-1.3.2-1.mga6.x86_64
- zstd-1.3.2-1.mga6.x86_64 

This should be tested by someone who uses btrfs

CC: (none) => jim

Mageia 6 :: x86_64

Packages installed cleanly.
Not a btrfs system.

CC: (none) => tarazed25

Installed these packages and all others associated with the desktop kernel 4.14.5 update on a 64-bit system on my HP Probook 6550b. (i3, 8GB, Intel graphics, Intel wifi) The idea was to simulate a user getting these updates all at once.

All packages installed cleanly, and everything seems to be functioning as it should. Common apps (Firefox, GIMP, vlc) all work.

CC: (none) => andrewsfarm

I use btrfs (and snapper) as root filesystem and can test the main commands of btrfs-progs like btrfs subvolume, btrfs balance, btrfs scrub.

But I never used btrfs and compression. Any suggestion how to test?


Ulrich

Created attachment 9834 [details]
CLI output of tests

Test on a x86_64 UEFI system, btrfs as root on LVM.

Added a new device to the root filesystem, 
btrfs balance and other tests,
run a snapper rollback,
reboot,
and deleted the new device online.

After all these aggressive tests the system is still alive and sound.

Ulrich

After installing the new nvidia340 driver, I updated the firmware and all other appropriate packages related to the 4.14.5 server kernel update, on 64-bit and 32-bit systems on the same hardware:

ASRock motherboard, AMD Athlon X2 7750 processor, 8GB RAM, nvidia 9800GT graphics, Atheros AR9485 wifi adapter.

And, I executed the "dracut -f" command on each system.

After the reboot, both systems seemed to be running fine. No regressions noted.

Created attachment 9836 [details]
CLI output of tests, zstd compression

Added zstd compression to test.

The mount option is # mount -o compression=zstd <device name> <path>.

Compressed and uncompressed data may coexist. At a glance, everything is fine. Though I would not recommend compression, because compression disables the No_COW attribute. In my case I have /home on a btrfs subvolume and No_COW is set for the akonadi and baloo database files.
cf. https://btrfs.wiki.kernel.org/index.php/Compression

So set btrfs compression at your own risk!

Best regards,
Ulrich

(In reply to Ulrich Beckmann from comment #4)
> I use btrfs (and snapper) as root filesystem and can test the main commands
> of btrfs-progs like btrfs subvolume, btrfs balance, btrfs scrub.
> 
> But I never used btrfs and compression. Any suggestion how to test?
> 
> 
> Ulrich

Some remarks

On reboot I had to chroot into the root partion and recreate the initrd and grub.cfg

# mkinitrd --force ...
# grub2-mkconfig -o ...

Before deleting the additional device the correct btrfs balance statement to restore the original state would be

# btrfs balance start --force -mconvert=dup -sconvert=dup /

I am ready now to set ok, go!

Ulrich


I recommend for btrfs beginners some finger exercises 
https://www.linux.com/learn/how-manage-btrfs-storage-pools-subvolumes-and-snapshots-linux-part-1
https://www.linux.com/learn/how-create-and-manage-btrfs-snapshots-and-rollbacks-linux-part-2

Whiteboard: (none) => MGA6-64-OK

Advisory added to svn:

type: bugfix
subject: Updated btrfs-progs packages adds support for kernel 4.14
src:
  6:
   core:
     - btrfs-progs-4.14-2.mga6
     - zstd-1.3.2-1.mga6
description: |
  Updated btrfs-progs adds support for new features found in 4.14 series
  kernels including zstd compresssion support.
references:
 - https://bugs.mageia.org/show_bug.cgi?id=22150

Keywords: (none) => advisory

The following 3 packages are going to be installed:

- btrfs-progs-4.14-2.mga6.i586
- libbtrfs0-4.14-2.mga6.i586
- libzstd1-1.3.2-1.mga6.i586

1.1MB of additional disk space will be used.

1016KB of packages will be retrieved.

Is it ok to continue?


/dev/sda7 on /home type btrfs (rw,relatime,space_cache,subvolid=5,subvol=/)

$ uname -a
Linux localhost 4.9.56-desktop-1.mga6 #1 SMP Thu Oct 12 22:53:48 UTC 2017 i686 i686 i686 GNU/Linux

Rebooted and confirmed file system is still active and in rw mode.

Confirmed samba still able to service requests.

Working as designed.

CC: (none) => brtians1
Whiteboard: MGA6-64-OK => MGA6-64-OK MGA6-32-OK

@Brian: thanks for your tests on this one. Validating for you.

Keywords: (none) => validated_update
CC: (none) => lewyssmith, sysadmin-bugs

An update for this issue has been pushed to the Mageia Updates repository.

https://advisories.mageia.org/MGAA-2017-0138.html

Status: NEW => RESOLVED
Resolution: (none) => FIXED