Previously, onionAddresses definitions in onionServices were of the form
onionAddresses.access.<service> = [<service>];
This caused failures for configurations where a service user name was
overridden or for bitcoind whose default user is 'bitcoin' instead of 'bitcoind'.
Now set the equivalent of:
onionAddresses.access.<actualServiceUser> = [<service>];
Implement this via a new option `onionAddresses.services` to make things more
readable and to work around an infinite recursion error in onionServices.
- bitcoind 0.20.1 -> 0.21.0
Manually create a wallet in the backup test because bitcoind
does not create a default wallet anymore
- disable the failing elementsd build on unstable
Only check enabled services. This allows running the backup test with a custom subset of other
tests.
Also, show a meaningful error on test failure.
Previously, just an AssertionError without a message was shown.
Because the test code is evaluated from a string and not a file, there
was also no backtrace to the tests file.
- Don't fail on `kill $cachixPid` when cachix has already exited.
This fixes some failing CI builds.
- Stop the cachix background worker before the final `cachix push`.
This can avoid unneeded reuploads.
Use the coreutils version of tail on cirrus which supports
argument --pid.
- enable usage outside of secure-node.nix
- use json as the output format
- show ports
- also show local addresses, which is particularly useful when
netns-isolation is enabled
- only show enabled services
Move 'enforceTor' and onion-service definitions from secure-node.nix.
Use the onionServices module to define onion services.
Onion services now automatically work for services that bind to an INADDR_ANY (`0.0.0.0`) address.
- use -e to check for existence of /var/lib/tor/state, use shorter
polling interval
- clear existing dataDir contents to avoid accumulating obsolete data
- use concatMapStrings instead of foldl'
The 'scudo' memory allocator set by the 'hardened' profile breaks some
services on 20.09.
The fix for NixOS unstable (https://github.com/NixOS/nixpkgs/pull/104052)
is ineffective on 20.09.
As a workaround, add a custom 'hardened' preset that uses the default allocator.
Password length and alphabet is unchanged, but the restriction to
include at least one numeric and one capital char has been removed.
This restriction is not needed by client applications,
adds code complexity, and even (insignificantly) reduces entropy.
Reason for switching to pwgen:
apg uses /dev/random instead of /dev/urandom which brings no security
benefits but can stall the generate-secrets script on low-entropy
devices due to blocking.
Since `security.rngd` has been disabled in NixOS 20.09, blocking
in generate-secrets can also appear on regular NixOS desktop systems.
Needed by the following commits.
Also, don't use the cachix cache as a substituter for local, non-CI
builds. This obviates the need for the 'untrusted' warning in build.sh.
Restore the original behavior that was accidentally changed:
When no args are given, run the basic test suite.
Otherwise, run the given command with default scenario 'default'.
Previously, `run-tests.sh build` ran the basic test suite instead of
building the default scenario.
- Make more economic use of the free CI resources by removing redundant build tasks:
- Build unstable pkgs in a single separate task ("pkgs_unstable").
- All stable pkgs are implicitly built by the modules tests.
- The build script (ci/build.sh) can now be executed locally for easier
debugging.
- Use an explicit 'cachix push' command instead of helper/wait-for-network-idle.rb.
This is simpler and more reliable.
Since nixpks 20.09, the test output is just an empty directory.
Restore saving the log output and linking to the driver.
Without linking to the driver, the driver is eligible for
garbage collection after running a test via `run-tests.sh --out-link-prefix`,
which implies lengthy driver rebuilds.
Under normal circumstances, service-specific netns should never exist
before the netns setup service starts.
An existing netns is a genuine error that should not be silently ignored.
A short time after `netns delete` finishes, the peer link in the main
netns is automatically removed.
When `link del` is run before that, it fails with
`Cannot find device "nb-veth-br-*"` and the netns service enters a failed state.
When 'loop getparams' fails, jq gets no stdin and exits with code 0.
Because -o pipefail is not enabled in the testing shell, the whole test
command succeeds, although it should fail.
Just test "loop getparams" instead and ignore its output.
This also enables the test scenario 'netnsRegtest' introduced in a
later commit by fixing the following bug:
For unknown reasons, when tor.active=true and tor is not running, lnd
fails with a tor connection error on netns-isolation, but runs fine
without netns-isolation.
- Use fping for pinging multiple hosts in parallel.
Significantly improves test runtime:
>13 s -> ~200 ms for the negative ping tests.
- Only test network namespaces that are enabled.
This allows running the netns test with a reduced service set for debugging.
- Remove deprecated services, instead add btcpayserver, spark-wallet
generate-secrets failed due to the `PYTHONPATH` env var that was set
through the nixops19_09 buildInput.
Fixes:
- Don't pull in the build environments of binaries that should be
available in PATH. Only extend PATH instead.
- Run generate-secrets in an empty environment
Erik Arvstedt