When adding a new virtual network interface (bond,bridge,vpn) or
configuring a new connection (e.g. ethernet/wifi) don't activate it
directly but allow a system administrator to configure it first.
This behaviour is similar to what GNOME does when adding new network
interfaces.
Closes: https://redhat.atlassian.net/browse/RHEL-34153
After creating a Bond/VPN/Team/VLan interface navigate to interface
details page so the user can further configure the interface itself.
In a follow up commit the connection will be disabled on creation so
navigating to it will make that more clear.
Let's not try to use machines with less RAM than the default, which is
a little bit more than 1G. Recent centos-10 images fail to boot with
less than 1G.
replace following properties on IP and IPv6 objects
addresses -> addressData + gateway
routes -> route-data
dns -> dns-data
dns-data is available since NM 1.42.0 so also keep support for older
dns prop as it is required on ubuntu-2204.
ipaddr.js is used to validate addresses so cockpit does not only rely on
network manager to handle this.
Huge team-effort to redesign Cockpit with Patternfly v6 in mind and to
reduce the amount of hacks and legacy code we have. This also paves the
way to further improve Patternfly itself by keeping up-to-date with
their latest releases.
Commit includes codemod changes, updated tests, improved layout for
desktop and mobile, as well as inline-comments here and there for what
we need to fix.
Co-authored-by: Garrett LeSage <garrett@redhat.com>
Co-authored-by: Tomas Matus <tomatus777@gmail.com>
Co-authored-by: Martin Pitt <mpitt@redhat.com>
Relates-to: #21651, #20973https://issues.redhat.com/browse/COCKPIT-1266
- Generalize the nginx skip (it's also not installed on
centos-9-stream, we just never have started to test it there on our
CI)
- vdo and wireguard are not currently available in RHEL 10
- NM team devices are not supported, see
https://issues.redhat.com/browse/RHEL-33026
Ubuntu ≥ 23.04 (mantic) integrates netplan into NetworkManager much more
tightly. The primary config is now in netplan. That particular check
just concerns our internal error handling of the networking page, and
it's enough to run it on one OS. Skip this on all Ubuntus, so that we
won't trip over this again once we move ubuntu-2204 to -2404.
Commit ccab549331 broke the checking of browser errors, as with an
unset `$TEST_ALLOW_JOURNAL_MESSAGES` (which is the case in our CI) it
was adding the empty string as allowed pattern, which always matches.
Ignore the React errors that happen with our current code, and add some
explanations. We have so many of them that we can't fix all of them in
one go, but we have to put a stop to introducing new ones. Some of them
are test specific.
Co-Authored-By: Martin Pitt <mpitt@redhat.com>
For some properties like `public-key` NM doesn't mind extra spaces, but
for others it might complain about passing invalid input. Fix it by
removing the spaces before sending it to NM.
Though the original issue #19384 used wg-quick to configure wireguard, I
was unable to reproduce the issue with it, unless I created a brand new
connection which seems unnecessary as the tests already had a profile
configured. Using `sed` to edit wg keyfile and reproduce should be
equivalent.
Other wireguard properties `endpoint` and `listen-port` are either
optional or have a default value, so the /network page shouldn't break
if they're not set. Missing or invalid `public-key` is equivalent to no
peer, so it shouldn't break /network page as well.
fixes#19384
