Re: btrfs-6.9 misc-tests.sh failures

[Qu Wenruo <quwenruo.btrfs@gmx.com>]

在 2024/6/11 05:08, Bruce Dubbs 写道:
> failed: /build/btrfs/btrfs-progs-v6.9/btrfs balance start
> -mconvert=single -sconvert=single -f
> /build/btrfs/btrfs-progs-v6.9/tests/mnt
> test failed for case 004-shrink-fs

This is a known bug and would be fixed by this patch:

https://lore.kernel.org/linux-btrfs/5a292583be11ae383e79aaca0fa79be2141ef6ca.1717732459.git.wqu@suse.com/T/#u

Thanks,
Qu
>
> The misc-tests-results.txt gives:
>
> ====== RUN CHECK /build/btrfs/btrfs-progs-v6.9/btrfs filesystem resize
> 8091860992 /build/btrfs/btrfs-progs-v6.9/tests/mnt
> Resize device id 1 (/dev/loop0) from 7.54GiB to 7.54GiB
> ====== RUN CHECK /build/btrfs/btrfs-progs-v6.9/btrfs balance start
> -mconvert=single -sconvert=single -f
> /build/btrfs/btrfs-progs-v6.9/tests/mnt
> ERROR: error during balancing '/build/btrfs/btrfs-progs-v6.9/tests/mnt':
> No space left on device
> There may be more info in syslog - try dmesg | tail
> failed: /build/btrfs/btrfs-progs-v6.9/btrfs balance start
> -mconvert=single -sconvert=single -f
> /build/btrfs/btrfs-progs-v6.9/tests/mnt
> test failed for case 004-shrink-fs
>
> How much space is needed?  On my /build partition:
>
> SOURCE         TARGET    FSTYPE     SIZE   USED AVAIL USE%
> /dev/nvme0n1p7 /build    ext4       245G 141.2G 91.3G  58%
>
> Disabling case 004-shrink-fs the next issue is
>
> ====== RUN MUSTFAIL /build/btrfs/btrfs-progs-v6.9/btrfs subvolume delete
> /build/btrfs/btrfs-progs-v6.9/tests/mnt/snap1
> Delete subvolume 257 (no-commit):
> '/build/btrfs/btrfs-progs-v6.9/tests/mnt/snap1'
> succeeded (unexpected!): /build/btrfs/btrfs-progs-v6.9/btrfs subvolume
> delete /build/btrfs/btrfs-progs-v6.9/tests/mnt/snap1
> unexpected success: deleting default subvolume by path succeeded
> test failed for case 041-subvolume-delete-during-send
>
> I do not have any insight into this issue.
>
> After disabling 041-subvolume-delete-during-send all other tests pass.
>
> -----
>
> Note, the reason we run tests in linuxfromscratch it to give users
> confidence in the packages built.  We can just disable some tests and
> label them as known errors,
> but it seems that an important package like btrfs should test clean.
> If the test failures are something we have done wrong, we would like to
> fix it.
>
>    -- Bruce Dubbs
>       linuxfromscratch.org
>