Your message dated Fri, 28 Mar 2025 04:19:28 +0000
with message-id <e1ty1bo-00gkgs...@fasolo.debian.org>
and subject line Bug#1101175: fixed in systemd 257.4-4
has caused the Debian Bug report #1101175,
regarding systemd: mounts a tmpfs on /tmp in the middle of upgrade
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact ow...@bugs.debian.org
immediately.)


-- 
1101175: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1101175
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems
--- Begin Message ---
Package: systemd
Version: 257.4-3
Severity: critical
Justification: breaks apt completely

I was upgrading my system from bookworm to trixie when it broke in the
middle with apt complaining about not finding a set of .debs in /tmp.
It turned out that systemd had activated the tmp.mount unit in the
middle of the upgrade.

As this was a large dist-upgrade, I'm not sure exactly what happened,
but some excerpts from the journal:

Mar 24 06:54:36 x systemd[1]: tmp.mount: Directory /tmp to mount over is not 
empty, mounting anyway.
Mar 24 06:54:36 x systemd[1]: Mounting tmp.mount - Temporary Directory /tmp...
Mar 24 06:54:36 x systemd[1]: Mounted tmp.mount - Temporary Directory /tmp.
Mar 24 06:54:36 x systemd[1]: Starting man-db.service - Daily man-db 
regeneration...
Mar 24 06:54:40 x systemd[1]: man-db.service: Deactivated successfully.
Mar 24 06:54:40 x systemd[1]: Finished man-db.service - Daily man-db 
regeneration.

man-db.service is a service with PrivateTmp=true, and according to
advice on #debian-devel, such a service will have Wants/After=tmp.mount.
The masking of the tmp.mount unit happens in postinst, so there's a race
where the new package might be unpacked, something calls systemctl
daemon-reload and then a service with PrivateTmp gets activated before
systemd's postinst is called.

I'm filing this as critical since it left the system completely broken
and required extensive use of dpkg -i to recover.

A better solution is probably to do the masking in preinst instead of
postinst, assuming my analysis is correct.

-- 
Tollef Fog Heen
UNIX is user friendly, it's just picky about who its friends are

--- End Message ---
--- Begin Message ---
Source: systemd
Source-Version: 257.4-4
Done: Luca Boccassi <bl...@debian.org>

We believe that the bug you reported is fixed in the latest version of
systemd, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 1101...@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Luca Boccassi <bl...@debian.org> (supplier of updated systemd package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmas...@ftp-master.debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Format: 1.8
Date: Fri, 28 Mar 2025 01:24:32 +0000
Source: systemd
Architecture: source
Version: 257.4-4
Distribution: unstable
Urgency: medium
Maintainer: Debian systemd Maintainers 
<pkg-systemd-maintain...@lists.alioth.debian.org>
Changed-By: Luca Boccassi <bl...@debian.org>
Closes: 1079329 1098914 1100653 1101175
Changes:
 systemd (257.4-4) unstable; urgency=medium
 .
   [ Luca Boccassi ]
   * Update tzdata-legacy version constraint for new bookworm-p-u tzdata
     version
   * systemd-boot: do case-insensitive parsing of efibootmgr output
     (Closes: #1100653)
   * userdbd: use wildcard to install units
   * systemd-container: use more wildcards to install files
   * Handle new tmp mount on upgrade in preinst rather than postinst. Avoid
     race with new units being activated and pulling it in at the same time
     (Closes: #1101175)
   * autopkgtest: remove workaround for 'meson compiile mkosi' and use
     'mkosi -f' instead to avoid slow rebuilds
   * autopkgtest: drop versioning constraints on tzdata. These do not seem
     to be necessary in order for autopkgtest to pull in the package, so
     drop them as they need to be constantly updated due to p-u uploads
   * Move leftover resolved files to its own package. A few files were
     mistakenly shipping in the main systemd package even though they are
     part of resolved, due to a refactor from some time ago. Move them
     over.
   * Drop systemd-resolved package. The ctte has declared that the way the
     systemd-resolved tool works is incompatible with their decision to
     prioritize avahi in Debian. Furthermore, the resolved tool is being
     used to inflict pain on the maintainer, and induce burnout.
     Regrettably, the only safe solution to ensure this package is
     compliant with this decision is to drop it, as all reasonable
     alternatives put forward have been rejected:
     https://salsa.debian.org/systemd-team/systemd/-/merge_requests/289
     (Closes: #1098914)
   * Drop systemd-nspawn on arm64. The ctte has declared that the way the
     systemd-nspawn binary works on arm64 is incompatible with their
     decision to leave in place a recently introduced change in the debian-
     specific base-files package, which caused a dpkg regression.
     Furthermore, the nspawn tool is being used to inflict pain on the
     maintainer, and induce burnout. Regrettably, the only safe solution to
     ensure this package is compliant with this decision is to drop the
     binary on arm64, as all reasonable alternatives put forward have been
     rejected:
     https://salsa.debian.org/systemd-team/systemd/-/merge_requests/289
     (Closes: #1079329)
 .
   [ Nick Rosbrook ]
   * Add myself to Uploaders
 .
   [ Andrea Pappacoda ]
   * d/control: only conflict against opensysusers << 0.7.3-4.1~
     opensysusers 0.7.3-4.1 is no longer co-installable with systemd as it
     conflicts with systemd-sysusers, so it is no longer needed to conflict
     against it systemd-side.
Checksums-Sha1:
 04f7adf6ca3a926c2cf8c28e5f712b7cd6c3fd35 8673 systemd_257.4-4.dsc
 ee29fff5205815634b2048521e76e73a4c602cb9 180508 systemd_257.4-4.debian.tar.xz
 64ded135987fe08c4befc986077c980fb40cf5c9 14031 systemd_257.4-4_source.buildinfo
Checksums-Sha256:
 8e485e45c018b2b8d97ad6dfaa4f062cc1b1b9c4919ebc180c93ce1b2e74e623 8673 
systemd_257.4-4.dsc
 975700ef87fc8ebe7b6c1b07fc33e0c40748532d7e5fcdb31f6628c766c71b6c 180508 
systemd_257.4-4.debian.tar.xz
 94df462382e965292995726371f8161838e5717cb84d352467842c27a59daacb 14031 
systemd_257.4-4_source.buildinfo
Files:
 652396c841c6db6223e3baef51050f1a 8673 admin optional systemd_257.4-4.dsc
 303a4daa0020340e8d317d74fa190e6b 180508 admin optional 
systemd_257.4-4.debian.tar.xz
 0e8aa0467a92d0968dde8563289da6b2 14031 admin optional 
systemd_257.4-4_source.buildinfo


-----BEGIN PGP SIGNATURE-----

iQJFBAEBCgAvFiEErCSqx93EIPGOymuRKGv37813JB4FAmfmHaERHGJsdWNhQGRl
Ymlhbi5vcmcACgkQKGv37813JB59hQ/9HRMT9tDP8Go3IaZIrWbEQMsGMSjqABqR
Y2sACH0qbdDw78o3XtlnJAP7Kja2llx5V6AP0asRz1TT5aoKch6p4S43pSwvcUR6
e5HwJ9giu2ueFqa9oiidkq/jMgp/JhqwfvXWJhtTRYwGIljezpPYwv+E/u1xB6Ri
Ff+6pdMcMewjFKz3of7o6ZYcO9eDpV5oi1fbPbLzUihGBQWKJKAY5svhqIKbwU+Y
4pc+dmTAj55xZYm5PRWo/c9jXIk4NebAPaqX8uhP7hBUux9DHlHUtpahn+a6iith
+6ORskBV9GBO0LDyfWOjcxvC0OO6Q8AkfRxtOza0KA/4ePE1OVmfiNHVhLVmgL7G
FcvgOqf3m0F/lFKKjL0mGtij+sGEmyh9F7VIEHWeDZxDslnunEk3v1qkf9ERsRvg
oAvOrmAG5okD0hwqwTNZWv7k2LKOEzwEj8F6lUDVfo0E+klcp+W/AtdqEA0rVEUE
aDcgiKJp/nknATGIQj2sLIMCN0jG6slYPyU9xUTTy+KArK6gOcfkEkwkmQ6fuANB
3QxshTwfTuIauv5Xo4vTVEfE+3LGC3QP0KTa3ZiRAORsUCyslXv3cyJdVhyalw8q
S9D013FLAK/wqMnwRs504pGVEqeq2h8iXPZFGp/RLgeAiHEXkKtunAKOl6FWVyBG
XQ4PZCcV4fE=
=Ni7s
-----END PGP SIGNATURE-----

Attachment: pgpRifeGNnwYz.pgp
Description: PGP signature


--- End Message ---

Reply via email to