Encrypted EXT4 fails on different devices on semi-fresh install: Delayed block allocation failed for inode

Support
1

My full-disk encrypted Manjaro corrupts storage for no apparent reason.

The reasons that I can think are ruled out:

  • hardware error (different notbook this time)
  • a specific program (no programs that are installed seem to influence the appearence)
  • wired configuration: I just followed the graphical installer and just selected encrypted with swap (ext4)

I also think this issue is Manjaro specific as Endeavour OS doesn’t show this behaviour.

Here is a short log from journalctl -xe:

Okt 23 13:40:24 falcon kernel: EXT4-fs (dm-0): Delayed block allocation failed for inode 47523228 at logical offset 0 with max blocks 2 with error 117
Okt 23 13:40:24 falcon kernel: EXT4-fs (dm-0): This should not happen!! Data will be lost
Okt 23 13:40:24 falcon kernel: EXT4-fs (dm-0): Delayed block allocation failed for inode 47523247 at logical offset 0 with max blocks 2 with error 117
Okt 23 13:40:24 falcon kernel: EXT4-fs (dm-0): This should not happen!! Data will be lost
Okt 23 13:40:24 falcon kernel: EXT4-fs (dm-0): Delayed block allocation failed for inode 47523249 at logical offset 0 with max blocks 3 with error 117
Okt 23 13:40:24 falcon kernel: EXT4-fs (dm-0): This should not happen!! Data will be lost

I already encountered this issue when last trying Manjaro - on another notebook: EXT4 Fileerrors after fresh install with different SSDs

i currently have a different notebook, fresh install. After ~1 year of running Endeavour OS without issue I wanted to try Manjaro again. Now two weeks in this happened again.

Here is another snippet from the logs

➜  ~ journalctl | grep EXT4
Sep 17 18:27:08 falcon kernel: EXT4-fs (dm-0): 13 orphan inodes deleted
Sep 17 18:27:08 falcon kernel: EXT4-fs (dm-0): recovery complete
Sep 17 18:27:08 falcon kernel: EXT4-fs (dm-0): mounted filesystem 3fb33050-7c92-47ce-98ed-83dfe8937301 r/w with ordered data mode. Quota mode: none.
Sep 17 18:27:08 falcon kernel: EXT4-fs (dm-0): unmounting filesystem 3fb33050-7c92-47ce-98ed-83dfe8937301.
Sep 17 18:27:08 falcon kernel: EXT4-fs (dm-0): mounted filesystem 3fb33050-7c92-47ce-98ed-83dfe8937301 r/w with ordered data mode. Quota mode: none.
Sep 17 18:27:08 falcon kernel: EXT4-fs (dm-0): re-mounted 3fb33050-7c92-47ce-98ed-83dfe8937301 r/w. Quota mode: none.
Sep 17 19:50:45 falcon kernel: EXT4-fs (dm-0): 8 orphan inodes deleted
Sep 17 19:50:45 falcon kernel: EXT4-fs (dm-0): recovery complete
Sep 17 19:50:45 falcon kernel: EXT4-fs (dm-0): mounted filesystem 3fb33050-7c92-47ce-98ed-83dfe8937301 r/w with ordered data mode. Quota mode: none.
Sep 17 19:50:45 falcon kernel: EXT4-fs (dm-0): unmounting filesystem 3fb33050-7c92-47ce-98ed-83dfe8937301.
Sep 17 19:50:45 falcon kernel: EXT4-fs (dm-0): mounted filesystem 3fb33050-7c92-47ce-98ed-83dfe8937301 r/w with ordered data mode. Quota mode: none.
Sep 17 19:50:45 falcon kernel: EXT4-fs (dm-0): re-mounted 3fb33050-7c92-47ce-98ed-83dfe8937301 r/w. Quota mode: none.

Info on my setup:

/etc/fstab:

UUID=2C2A-850C                            /boot/efi      vfat    umask=0077 0 2
/dev/mapper/luks-318ef27f-7c03-4741-b62e-a98636c5deef /              ext4    defaults,noatime 0 1
/dev/mapper/luks-87489635-e8a3-421a-9b70-2a84c33da56a swap           swap    defaults,noatime 0 0
tmpfs                                     /tmp           tmpfs   defaults,noatime,mode=1777 0 0

lsblk -l

NAME                                      MAJ:MIN RM   SIZE RO TYPE  MOUNTPOINTS
luks-318ef27f-7c03-4741-b62e-a98636c5deef 254:0    0 919,8G  0 crypt /
luks-87489635-e8a3-421a-9b70-2a84c33da56a 254:1    0  33,8G  0 crypt [SWAP]
nvme0n1                                   259:0    0 953,9G  0 disk  
nvme0n1p1                                 259:1    0   300M  0 part  /boot/efi
nvme0n1p2                                 259:2    0 919,8G  0 part  
nvme0n1p3                                 259:3    0  33,8G  0 part

I am happy for any ideas how to debug this.

2

Maybe these recent changes to default password hashing algorithm and umask settings are related.

1 Like
3

I don’t think as this problem is more than two years old and the changes seem to be recent…

4

My issue sadly isn’t solved and as I need the PC I switched to EndeavourOS and restored from backup. Sorry that I can’t help more to debug this but the issue is above my paygrade

5

The changes are indeed recent; thus the reasoning behind my stating that they were recent changes.

:point_up_2: This would also indicate to most people with a reasonably adequate command of the English language that your issue is recent.

My take:- If you haven’t a clue, then don’t be so eager to dismiss a suggestion simply because the concept might be deeper than the loose change in your pocket.

Enjoy the other OS. Cheers.

6

I don’t think your tone here is necessary. The point I was trying to make is that the timeline of the changes you pointed out does not seem to indicate any relation to this problem as I encountered the problem also two years ago. I didn’t want to be dismissive of it completely but I currently seem to not understand how the issue and the change you pointed out correlate. I am honestly thankful for the suggestion!

Also I didn’t want to criticize Manjaro in general, I think it’s a very nice OS and I’d love to be able to use it. wired bugs happen. I just wanted to communicate that I’m not able to debug this further (so future visitors of this thread know the state of this thread). I also wanted to point out that I use Endeavour OS because it’s also arch-based, so in my (limited) understanding quite similar and it might help the pinpoint the issue.

7

If only by merit of using Luks; and that failing; the referenced topic was likely related. Your problem of 1-2 years ago may have been different altogether, despite it seeming to be the same as you saw it.

Stating “two weeks in” as you did indicates that it worked for two weeks (until then it didn’t); again, the timeframe seems consistent with that of the information given in the link.

Is this the cause? I don’t know, specifically. However it seems a likely possibility that it at least contributed to your issue, considering it’s a change in the way authentication is handled by Manjaro; and likely extends to authentication of encrypted drives.

Some research might have revealed any direct relevance; but that’s your job, not mine.

Regarding tone; indeed, sometimes it’s necesary to shake preconceived assumptions and promote a constuctive outcome. Deal with people long enough in a forum environment, and you too would eventually see that. :slight_smile:

Thank you for taking the time for further clarification.

Cheers.