[SCW C1] Failed to start Network Name Resolution

[WarheadsSE]

Go to your node, and set the bootscript to `armv7l 4.1.22 std #1`, at least with this I could manage to get a console and an assh session back. Anything newer and I couldn't get that.

[WarheadsSE]

journald: status=228/SECCOMP
networkd: status=228/SECCOMP
resolved: status=1/FAILURE (above)

I have a feeling this is a a case across the board.

[Chazy]

If you mean the server won't boot anymore after a full system upgrade (no ssh session), take a look at https://github.com/scaleway/image-archl ... -229883103 and https://github.com/scaleway/image-archl ... -236698766 those fix the boot. I can use the latest kernel available (4.5.7) with these fixes.

I'm going to try this fix https://github.com/systemd/systemd/issu ... -239595447

They should be releasing a new image anytime soon. At least I hope that


EDIT: I tried a new C1 server in the following order and it worked, no more errors at boot:

1. Edit /etc/systemd/network/eth0.network (fixes boot)
2. Edit /etc/systemd/system/systemd-resolved.service (fixes systemd-resolved)
3. Full system upgrade
4. Delete and regenerate the machine id (fixes some errors)
5. Reboot

Let's hope it's fixed for good this time

P.S.: Is it just me, or did you too noticed that when installing a package (for example apache) it stays too long doing `(1/1) Updating manpage index...` before showing the prompt? I think it may be ssh lag or something...

[necbot]

I tried that fix you linked to on github. It doesn't work. This is a very frustrating bug. I have a new installation of Arch ARM but I can't update/install or do anything. Since systemd-resolved service fails to start DNS doesn't work. Every pacman command I run has the same error "cannot resolve mirror.archlinuxarm.org". This new installation of Arch ARM is essentially worthless. How the heck did this bug even make it into the new installation image??? I'm about to throw in the towel and switch to ubuntu ARM.

[WarheadsSE]

https://github.com/scaleway/image-archl ... -239679800

[WarheadsSE]

And that "too long" is because it is flushing the write cache with a `sync` call before returning. The entire rootfs is on NFS, after all.

[WarheadsSE]

1) regenrate machine-id
$this->bbcode_second_pass_code('', '
# rm /etc/machine-id
# systemd-machine-id-setup
')
2) use a proper drop-in for systemd-resolved
$this->bbcode_second_pass_code('', '
# cd /etc/systemd/system
# rm systemd-resolved.service
# mkdir systemd-resolved.service.d
# echo "[Unit]" >> systemd-resolved.service.d/10-networkmanager-wait-online.conf
# echo "After=NetworkManager-wait-online.service" >> systemd-resolved.service.d/10-networkmanager-wait-online.conf
')
3) Check your eth0.network
$this->bbcode_second_pass_code('', '
# cd /etc/systemd/network
# cat eth0.network
')
Should read this at the bottom. If not, add it.
$this->bbcode_second_pass_code('', '
[DHCP]
CriticalConnection=true
')

[WarheadsSE]

$this->bbcode_second_pass_code('', '
$ ssh root@212.47.231.205
Last login: Sun Aug 14 15:57:26 2016 from 174.54.190.115
[root@scw-405265 ~]# pacman -Syu
:: Synchronizing package databases...
core is up to date
extra is up to date
community is up to date
alarm is up to date
aur is up to date
:: Starting full system upgrade...
there is nothing to do
[root@scw-405265 ~]# uptime
16:06:57 up 0 min, 1 user, load average: 0.24, 0.06, 0.02
[root@scw-405265 ~]# date
Sun Aug 14 16:06:58 UTC 2016
[root@scw-405265 ~]#
')

[necbot]

my systemd-resolved.service file is in...

/etc/systemd/system/multi-user.target.wants/systemd-resolved.service

not

/etc/systemd/system/systemd-resolved.service

does this make a difference?

[WarheadsSE]

/me SMH

The actual file? or a symlink?