This is useful when the user-data needs to reboot an instance.
Previously, some initialization did not happen which is now being done
first before the user-data script is being executed.
This gives users more flexibility about what they are doing in those
scripts.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Terraform only supports sending any shell scripts encoded in base64
which is however not required by Oracle. Therefore we have to test if
the script is encoded or not.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Reviewed-by: Peter Müller <peter.mueller@ipfire.org>
Bumping across one of our scripts with very long trailing whitespaces, I
thought it might be a good idea to clean these up. Doing so, some
missing or inconsistent licence headers were fixed.
There is no need in shipping all these files en bloc, as their
functionality won't change.
Signed-off-by: Peter Müller <peter.mueller@ipfire.org>
Fix for Bug 12445: There is no proper way to distinguish between MS Azure and a local
Hyper-V installation Procedure: try to get the metadata info from the
internal MS metadata service at IP 169.254.169.254 If this fails the system
is running on an local Hyper-V instance wget takes too long to exit if the IP
is not reachable Added --timeout and --tries param to wget to reduce the
duation from 8 minutes to 9 seconds 9 seconds at one try should be enough
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Signed-off-by: Thomas Cekal <admin@cekal.org>
This variable is no longer being used and was only used to
assign IP addresses to the individual interfaces.
However, the kernel knows best which IP address to select
as broadcast address for each network. Therefore we depend
on the kernel which allows us to support RFC3021.
Fixes: #12486 - no /31 transfer net available on red
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
The whole hostname was used as domain name because there
was no . in it where the string could have been split.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
AWS supports jumbo-frames which IPFire can take advantage of
to increase network throughput internally.
The MTU for RED was left as 1500 to avoid packet fragmentation
in the cloud network and have IPFire do that job.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
This is mostly aesthetic because there are no ISP nameservers
anyways that we could use here.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Signed-off-by: Arne Fitzenreiter <arne_f@ipfire.org>
We cannot reliably determine if a system is running on Hyper-V
on a private server or on the Azure Cloud.
Therefore, we will have to try to retrieve an IP address
with DHCP and try to connect to the metadata service. If either
of those things is not successful, we will just continue with
the setup process as usual.
So cloud instances should be automatically configured now and
all other systems will continue to boot and call the setup
wizard as usual.
Fixes: #12272
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Signed-off-by: Arne Fitzenreiter <arne_f@ipfire.org>
Those scripts used to import settings from the meta-data services
and wrote them to the local configuration files.
For the DNS settings and Amazon, this is no longer possible because
their DNS servers do not support DNSSEC at all. Therefore we default
to recursor mode.
To be consistent across cloud providers, we are doing the same for
Azure.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Signed-off-by: Arne Fitzenreiter <arne_f@ipfire.org>
sshctrl calls sshd directly which won't work at time of the first boot
because no keys will be generated.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Amazon does not permit that a user logs in as root directly.
Instead they insist on using sudo.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
