#798887 btrfs-progs: btrfs filesystem defrag order of arguments different results

#798887#5
Date:
2015-09-13 20:14:44 UTC
From:
To:
Dear Maintainer,


I use the show-shared-extents (duperemove) tool from markfasheh v0.10 and
filefrag to check on the extents.
Before I used it I ran "btrfs filesystem defrag -clzo". Then a second time
"btrfs filesystem defrag -v -clzo". And also "btrfs filesystem defrag"

I noticed three things.
* without the -clzo flags I noticed change in extents, does this uncompress the
files even the man page says otherwise
* -v -clzo and -clzo -v seems to produce different number of extents, could it
be that -v -clzo does same as "btrfs filesystem defrag"
* when I did btrfs filesystem defrag -czlib and then after that with -clzo two
times extent size changed in the second run, when I expected it only to change
in the first run. (after that the show-shared-extents reports fe_length: 524288
for all but the last extent, after the first run some extents had different
size)

#798887#8
Date:
2015-09-13 20:26:32 UTC
From:
To:
Control: reassign -1 btrfs-tools
#798887#19
Date:
2025-10-11 22:14:12 UTC
From:
To:
Hi,

Sorry no one replied to your bug sooner.  I became the maintainer of
btrfs-progs for Debian at 6.12-1.

guni <guni@anub.de> writes:

What version are you reporting for and can you still reproduce this?

Regards,
Nicholas

#798887#26
Date:
2025-10-12 01:07:28 UTC
From:
To:
Hmm, the submitter is no longer able to receive mail.

  > This message was created automatically by mail delivery software.

  > A message that you sent could not be delivered to one or more of its
  > recipients. This is a permanent error. The following address(es) failed:

  >   guni@anub.de
  >     host mail.anub.de [REDACTED]
  >     SMTP error from remote mail server after RCPT TO:<guni@anub.de>:
  >     550 5.2.2 <guni@anub.de>... Mailbox full  This message was created automatically by mail delivery software.

Guni, if ever you check up on this bug, this is one reason I closed it.
The other is because in 6.14 (maybe earlier) the manpage reads:

    -v     (deprecated) alias for global -v option

I confirmed the bug is still present on bookworm, and that '-clzo -v'
differs from '-v -clzo', but if it's deprecated in bookworm, then
upstream is already on its way to removing the cause of the bug.

Best,
Nicholas