I use a shell script unmodified for years to wakeup my nas. It failed recently. I checked I could still wakeup using android app, so I suspected my script. If basically does: wakeonlan 'NAS eth mac address' tested it as user and root. Not working anymore. replaced by etherwake 'NAS eth mac address' and it was working again. So I guess the wakeonlan perl script is somehow broken.
On Sun, 31 Aug 2025 15:41:28 +0200 Eric Valette <eric.valette@free.fr> wrote: Note that I tested upstream git current version and that it does not work either.
When was the last time it worked? When was the first time it did no longer work? Related to that is the question whether this is a regression in trixie or some breakage in unstable after trixie. cu Adrian
Just tested today after your message to be sure. It works again!!! No clue what change on my system corrected it (except I update my system everyday and change the upstream kernel for the last longterm one). I do not suspect kernel as etherwake was working at the same time wakeonlan was not.
FYI, the wakeonlan script hasn't changed since at least bookworm (and this may be much more longer, as the script says 2005, though I'm wondering whether this is up-to-date). I've noticed at my lab that packets seem to be sometimes randomly filtered or ignored. So it may be just luck that etherwake was working at that time. Unless the difference of behavior is reproducible, you cannot deduce anything. And if you can reproduce this, you should look at the differences with wireshark. Otherwise I think that this bug should be closed and the wakeonlan package should reenter testing (it has just been removed from testing).
The script itself surely not, but perl, perl packages, kernel, ... So this is not a sufficient reason. and almost no traffic (as I use it before starting transfer to my nas) . And again this was systematic : at a certain period of time wakeonlan did not send the magic packet or a corrupted version of it (did not run wireshark to see exactly). No error as program return code. And etherwake was systematically working. I did maybe ten of each and tried wakeonlan upstream version also. But as I cannot reproduce the bug anaymore and wakeonlan and etherwake both now work as expected I do not mind if the bug is closed. But since debian take the code upstream, upstream did some changes, and debian version is not up-to-date so I would take the bug as an opportunity to update the version to current upstream...
The script does some output:
print "Sending magic packet to $ipaddr:$port with $hwaddr\n";
and you can check that this is correct. So this would be a major
bug somewhere in basic networking, not in the script itself.
You can still check whether there are differences, even if this
is working. For instance, perhaps the default broadcast address
is different in wakeonlan and etherwake (this might matter in
some cases).
We believe that the bug you reported is fixed in the latest version of
wakeonlan, 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 1112642@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.
Debian distribution maintenance software
pp.
Petter Reinholdtsen <pere@debian.org> (supplier of updated wakeonlan 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 ftpmaster@ftp-master.debian.org)
Format: 1.8
Date: Sat, 27 Dec 2025 18:55:44 +0100
Source: wakeonlan
Architecture: source
Version: 0.42-1
Distribution: unstable
Urgency: medium
Maintainer: Thijs Kinkhorst <thijs@debian.org>
Changed-By: Petter Reinholdtsen <pere@debian.org>
Closes: 1003381 1112642
Changes:
wakeonlan (0.42-1) unstable; urgency=medium
.
* Team upload, salsa debian namespace.
.
* Added d/gbp.conf to enforce use of pristine-tar.
* New upstream version 0.42 (Closes: #1003381).
- Program should be working again (Closes: #1112642).
- Adjusted d/docs with new markdown README.md name.
- Refreshed 01-etcethers patch for new upstream.
* Updated Standards-Version from 4.1.5 to 4.7.3.
* Marked perl script binary package wakeonlan as Multi-Arch: foreign.
Checksums-Sha1:
a6f2fdea1e433679e2bffa33da99914fd1667d99 1826 wakeonlan_0.42-1.dsc
ba345f8cab220a02884a2c31a5d440ffb4617fe0 8981 wakeonlan_0.42.orig.tar.gz
cf63c3c50d5fed85f8822a019a1c908749a92947 4612 wakeonlan_0.42-1.debian.tar.xz
11843a096313b4b1090d7dd4969f2593bdbf7c06 5931 wakeonlan_0.42-1_source.buildinfo
Checksums-Sha256:
c83f07b6048aa5551169f592c4061747f2dfab850031689411834ba835ae8c7c 1826 wakeonlan_0.42-1.dsc
4f533f109f7f4294f6452b73227e2ce4d2aa81091cf6ae1f4fa2f87bad04a031 8981 wakeonlan_0.42.orig.tar.gz
897b577ece25aaa64f4d04c33c3574decacc80b44c3669563724eeb4f9d91a4b 4612 wakeonlan_0.42-1.debian.tar.xz
d6a21c3f050d18a4987721151ca876d26e79ce2b1af092456efa8389db0a9a48 5931 wakeonlan_0.42-1_source.buildinfo
Files:
045f15782acbf819208ee726020d8ef9 1826 net optional wakeonlan_0.42-1.dsc
12503e08f944dc0eff0971c7f1a27b04 8981 net optional wakeonlan_0.42.orig.tar.gz
b3adf0b9ff5c29ff001e328c9e99c4a6 4612 net optional wakeonlan_0.42-1.debian.tar.xz
e53c8cd0684fefe26ffacf31d8879e0f 5931 net optional wakeonlan_0.42-1_source.buildinfo
-----BEGIN PGP SIGNATURE-----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=pdCR
-----END PGP SIGNATURE-----
Hi, Given the initial report mentioned kernel 6.12.44 to be used this might just have been https://lore.kernel.org/regressions/20250822165231.4353-4-bacs@librecast.net/ which was for the 6.12.y series introduced in 6.12.43 and then fixed with 6.12.45. Regards, Salvatore
Thanks for the info. I really dislike bugs that disappear as they tend to reappear at the worst moment. However this means that Eterwake and Wakeonlan use differents means to send the WOL packet. That confused me. I did not supect the kernel as etherwake was working. Thanks anyway.