** Description changed:
- A customer reported network disconnections on their storage
+ [impact]
+
+ manual disabling of ipv4 DAD (IACD) for static link-local address does
+ not work in jammy
+
+ [test case]
+
+ see 'Reproducer' in original description below
+
+ [regression potential]
+
+
** Changed in: systemd
Status: New => Fix Released
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1964494
Title:
Setting DuplicateAddressDetection=none doesn't disable DAD for link-
local
** Changed in: systemd
Status: Unknown => New
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1964494
Title:
Setting DuplicateAddressDetection=none doesn't disable DAD for link-
local IPs
There is now an upstream fix to keep ACD on by default on ipv4ll
addresses, but still honor the user settings if they are set:
https://github.com/systemd/systemd/pull/22824
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
Thank you for reporting this. I've forwarded it to the upstream
developers, as I think at very least this new behavior should be
documented in
https://systemd.network/systemd.network.html#DuplicateAddressDetection=
https://github.com/systemd/systemd/issues/22763
** Bug watch added:
Status changed to 'Confirmed' because the bug affects multiple users.
** Changed in: systemd (Ubuntu)
Status: New => Confirmed
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1964494
Title:
I agree, a user-configured setting should be honored. Another thing to
consider is that according to the Jammy systemd.network man page, the
default should be 'ipv6' so one would expect that to be a global
default, but it gets overridden for link-local IPs due to this commit.
[1]
> I think this commit [1] may be related (landed in systemd v249)
yeah, that commit seems to be intentionally forcing ACD on for
statically configured ipv4, i'm not sure quite why, it seems like a
user-configured setting should be honored; but possibly Yu was trying to
default it to on instead of