unofficial mirror of bug-guix@gnu.org 
 help / color / mirror / code / Atom feed
From: Tobias Geerinckx-Rice via Bug reports for GNU Guix <bug-guix@gnu.org>
To: 40447@debbugs.gnu.org
Subject: bug#40447:
Date: Sun, 05 Apr 2020 18:01:32 +0200	[thread overview]
Message-ID: <87eet13orn.fsf@nckx> (raw)
In-Reply-To: <baea9ea5-a8e8-b0bd-8981-72c6101899da@brendan.scot>

[-- Attachment #1: Type: text/plain, Size: 1630 bytes --]

Brendan,

Brendan Tildesley 写道:
> Closing since it seems to unrelated. Sorry for wasting your 
> time. I
> think it may have been my ext4 having too many files. i deleted 
> all my
> system generations and gc'd a lot of stuff, and it seems to be 
> miss
> behaving. perhaps it is a different kind of guix bug that it 
> doesnt
> cope with this issue:

Worse: it's an ext4 bug (well, behaviour).

> [ 7862.196231] EXT4-fs warning (device sdb1): 
> ext4_dx_add_entry:2343:
> Directory (ino: 24903688) index full, reach max htree level :2

This happens when a directory, almost certainly /gnu/store/.links, 
contains more entries than can fit in ext4's ‘dir_index’ hash 
table.  The file system simply punts.

You can disable this optimisation with tune2fs -O ^dir_index …. 
You'll lose some theoretical performance but I haven't heard 
complaints in practice.  You're not the first one to encounter 
this problem.

If that's unacceptable, you can reconfigure your guix-daemon with 
‘--disable-deduplication’ in exchange for more storage space for 
newer generations.

Or switch to btrfs for a newer, shinier set of pathological 
gotchas but not this one :-)

One could work around this in Guix if one were motivated:
- Ignore file system failure instead of making it fatal, falling 
  back to ‘--disable-deduplication’ behaviour, which is probably 
  the right thing to do; or
- store .links in ./n/e/sted directories, which will degrade 
  performance on file systems that handle collisions properly and 
  is probably not worth the headache.

Kind regards,

T G-R

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]

  reply	other threads:[~2020-04-05 16:02 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-05 12:26 bug#40447: guix pull failure on master Brendan Tildesley
     [not found] ` <handler.40447.B.15860896423569.ack@debbugs.gnu.org>
2020-04-05 13:24   ` bug#40447: Acknowledgement (guix pull failure on master.) Brendan Tildesley
2020-04-05 14:39     ` Danny Milosavljevic
2020-04-05 14:46       ` bug#40447: Brendan Tildesley
2020-04-05 16:01         ` Tobias Geerinckx-Rice via Bug reports for GNU Guix [this message]
2020-04-05 17:23         ` bug#40447: Danny Milosavljevic
2020-04-05 14:39     ` bug#40447: Acknowledgement (guix pull failure on master.) Brendan Tildesley
2020-04-06  9:15 ` bug#40447: guix pull failure on master Ludovic Courtès

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://guix.gnu.org/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87eet13orn.fsf@nckx \
    --to=bug-guix@gnu.org \
    --cc=40447@debbugs.gnu.org \
    --cc=me@tobias.gr \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/guix.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).