* outsource Rspamd ENVs into explicit helper
This will allow us to uniformly source the helper and get the values
from everywhere consistently. This is more than desirable since we will
be using these values not only for the Rspamd setup, but also for DKIM
management and during change-detection.
* integrate Rspamd into changedetection
We outsource one more function to reside in the helper script for Rspamd
so that we can call this function from the Rspamd setup and from the
changedetection functionality too.
* realize deprecation of old commands file for Rspamd
THIS IS A BREAKING CHANGE!
This change realizes the log message: "Using old file location now
(deprecated) - this will prevent startup in v13.0.0" Startup will now
fail.
* added '--force' option to Rspamd DKIM script
* use new helper to get ENVs for Rspamd in DKIM script
* remove the need for linking directories
This was unnecessary, as explained in
https://github.com/docker-mailserver/docker-mailserver/pull/3597#discussion_r1369413599
* Apply suggestions from code review
review by @polarathene
* apply more review feedback from @polarathene
- <https://github.com/docker-mailserver/docker-mailserver/pull/3599#discussion_r1370885519>
- <https://github.com/docker-mailserver/docker-mailserver/pull/3599#discussion_r1370904201>
* update documentation
---------
Co-authored-by: Brennan Kinney <5098581+polarathene@users.noreply.github.com>
Instead of using `etc/rspamd/override.d/dkim_signing.conf`, we will now
be using `/tmp/docker-mailserver/rspamd/override.d/dkim_signing.conf`.
The new location is persisted (and linked again during startup) and
hence better suited.
* chore: Adjust default DKIM size (`open-dkim`) from 4096-bit to 2048-bit
4096-bit is excessive in size for DKIM key. 2048-bit is plenty.
* chore: Additional revisions to `open-dkim` command help output
- The examples use `keysize 2048`, but as that's the new default it makes sense to change that.
- Other help text was also revised.
- Last example for domains did not need to demonstrate the other options. Changed example domains to more appropriate values.
* docs: Revise DKIM docs
Primarily for the change in default key size, but does revise some text to better communicate to the user.
- While the referenced RFC advises 512-bit to 2048-bit key size, we now explicitly discourage `512-bit` as it's not secure. `1024-bit` is still likely safe for most, but `2048-bit` is a good default for those not rotating their keys.
- Adjusted the domains example to match the new `setup config dkim domain` domains example.
- Tip for changing default key size changed to "info" with added clarity of lowering security or increasing it (excessively).
- Rspamd section is minor formatting changes, with the exception of clarifying the "main domain" for the mail accounts is assumed as the DMS FQDN with any subdomain (like `mail.`) stripped away. This is not great, but a legacy issue that needs to be addressed in future.
- `docs-rspamd-override-d` ref removed, and usage replaced with equivalent ref `docs-rspamd-config-dropin`, while `docs-rspamd-config-declarative` ref was not in use and also removed.
- Revised the `<selector>.txt` DNS formatting info section to better communicate with the reader. Additionally it had mixed usage of default `mail` and custom `dkim-rsa` selectors (_file content and output_).
* docs: Sync DKIM commands help messages and update DKIM docs for LDAP
- Adopt the help options format style from the `rspamd-dkim` into `open-dkim` command. And convert `./setup.sh` to `setup`. `selector` option has been implemented. for a while now.
- Update `rspamd-dkim` examples help output to align with `open-dkim` command examples.
- Give both DKIM command tools a consistent description. The two tools differ in support for the `domain` option (_implicit domain sourcing for default account provisioner, and support for multiple domains as input_).
- DKIM docs for LDAP domain support revised to better communicate when explicit domain config is necessary.
* tests: Adjust test-cases for `setup config dkim` change
`rspamd_dkim.bats`:
- Update assert for command help output.
- Don't bother creating a DKIM key at 512-bit size.
`setup_cli.bats`:
- Update assert for command help output of the `setup config dkim` (OpenDKIM) command.
* docs: Update DKIM section for large keys to newer RFC
The linked discussion from 2021 does mention this updated RFC over the original. That removes outdated advice about `512-bit` key length support.
The discussion link is still kept to reference a comment for the reader to better understand the security strength of 2048-bit RSA keys and why larger keys are not worthwhile, especially for DKIM.
* docs: Extract out common DKIM generation command from content tabs
Should be fine to be DRY here, not specific to `open-dkim` or `rspamd` generation/support. Previously rspamd lacked support of an equivalent command in DMS.
* docs: DKIM refactoring
- Shifted out the info admonition on key size advice out of the content tabs as it's now generic information.
- Indented the 4096-bit warning into this, which is less of a concern as the default for our DKIM generation tools is consistently 2048-bit now.
- Reworked the LDAP and Rspamd multi-domain advice. To avoid causing a bad diff, these sections haven't been moved/merged yet.
* docs: Revise DKIM docs
Advice for managing domains individually with LDAP and Rspamd extracted out of the content tabs. Default domain behaviour explained with extra info about OpenDKIM + FILE provisioner sourcing extra domains implicitly.
* adjust learning of ham
See #3333
When moving a mail from the Junk folder to the Trash folder, the mail
previously classified as ham due to the wildcard match of `*`. Because
the syntax does not allow for negation, we can only change the behavior
in a way that mails are learned as ham when they are moved into `INBOX`
from `Junk`. This is reasonable though.
* adjust tests accordingly
* adjust docs accordingly
* add sanity check for Clam size & adjusted MaxScanSize
The second part is of special importance! See
<https://askubuntu.com/a/1448525>, which explains that the maximum scan
size is important as well. We previously just set the maximum file size,
which actually is pretty insecure as we silently not scan mile bigger
than `MaxScanSize`. This is corrected now.
* add SlamAV size configuration to Rspamd
* move modules adjustment file to new location
Because we link `/tmp/docker-mailserver/rspamd/override.d` to
`/etc/rspamd/override.d`, I think it makes sense to move the modules
adjustment file into `/tmp/docker-mailserver/rspamd/` as well.
I write the code in a way that it is backwards compatible for now, so
this is NOT a breaking change.
* minor improvement to `__rspamd__handle_user_modules_adjustments`
The expansion of `ARGUMENT3` is now done in a way that only adds the
whitespace in case the variable is set and not null.
* move test file structure to respect latest changes
Because we're now linking `rspamd/override.d/`, we can simplify the
setup a bit. But this requires a change in directory structure.
The current Rspamd test will be renamed to `rspamd_full.bats`, because I
plan on adding more tests in different files for different feature sets.
This is done to make this feature well-tested!
* improved and added tests to Rspamd-full
FYI: The line
```bats
_run_in_container grep 'sieve_global_extensions.*\+vnd\.dovecot\.pipe'
"${SIEVE_CONFIG_FILE}"
```
was testing a condition that should actually not be met, but when I
started working on this feature, I thought this was the correct
configuration. Adding the `assert_success` statements revealed this
wrong line.
I also added tests to check whether `override.d` is linked correctly.
* renamed: `rspamd.bats` => `rspamd_full.bats`
* added new tests for incomplete Rspamd feature set
We now test that warnings are emitted & features are disabled correctly.
* update documentation
* added checks whether OpenDKIM/OpenDMARC/policyd-spf are enabled
* added functions to check if VAR is 0/0 or an int
and also added tests.
I also adjusted the test file to not run in a container, because there
is no need. This also decreases test time, which, in turn, increases
maintainers' happiness.
* added more checks to Rspamd setup
I added the helpers from the previous commit to the Rspamd setup to make
the whole setup more robust, and indicate to the user that an ENV
variable's value is incorrect.
While we did not issues for this in the past, I believe it to be
worthwhile for the future.
* added canonical directory for users to place files in
This dir is canonical with DMS's optional configuration dirs, as it
lives in well-known volume mounts. Hence, users will not need to adjust
`/etc/rspamd/override.d` manually anymore, or mount a volume to this
place.
The docs explain this now, but the DKIM page needs a slight update on
this too I guess. I will follow-up here.
* misc minor improvements
* use variables for common directories
* postfix: remove smtpd_sasl_auth_enable global setting
* tests: disable auth on 25 port
* tests: revert ldap-smtp-auth-spoofed-sender-with-filter-exception.txt
* Skip failing test
The test seems to have been broken from the beginning.
Sadly, no LDAP maintainers can verify. Added a TODO item if ever a LDAP maintainer comes around.
* Apply PR feedback
---------
Co-authored-by: Georg Lauterbach <44545919+georglauterbach@users.noreply.github.com>
* refactor `mail_pop3.bats`
* refactor `mail_with_imap.bats`
* refactor `mail_with_relays.bats`
* moved test that that did not belong into POP3 test
* slightly clean up `no_container.bats`
Co-authored-by: Brennan Kinney <5098581+polarathene@users.noreply.github.com>
* add functionality for filtering mail log by ID
This was not planned, but as @polarthene mentioned in
https://github.com/docker-mailserver/docker-mailserver/pull/3033#issuecomment-1407169569
, filtering the mail log by email ID would be (the only) correct
approach for the Rspamd test (to eliminate race conditions).
I asserted the currect state, and came to the conclusion that this might
(or actually is) something we want in more than one place. So I went
ahead and implemented a solution.
The solution for acquiring the ID is a bit slower because it ensures the
mail queue is empty _before_ and _after_ the mail is sent. This is the
tradeoff one has to make if they want to send multiple emails in one
test file and get their IDs.
I hope you like this approach. I will provide another PR that adjusts
our current tests to use these new functions.
* added note about our helper functions in the docs
I think our work for our custom test framework should be noted in the
docs for newcomers to better understand what they should do.
* adjust Rspamd test to use new helpers for sending
* improve filter helpers further
* add sanity check when acquiring mail ID
* re-add `refute_output` to test which should now work well
* tests(fix): `spam_junk_folder.bats` Wait on Amavis port to be ready
Postfix can potentially be ready before Amavis is. This caused test failures as mail was sent before Amavis was ready to process it.
Both test cases shared the same test logic, except for the expected location to deliver the spam to. Extracted into a shared test method, and moved the port conditions into there.
* tests(chore): `spam_junk_folder.bats` minor revisions
Test case descriptions, container names and test prefix are now more descriptive of what is under test here (an ENV for Amavis).
* tests(chore): Move Amavis bounce test into `spam_junk_folder.bats`
These two tests seem to be related to the same feature. Grouping them into a single test file instead.
* tests(refactor): Split shared method into smaller methods
Now it can be better shared with the bounce test case.
* tests(chore): Shift test cases to match their CONTAINER_NAME order
No changes to code, just cut + paste of the `CONTAINER3_NAME` test case to shift it to the last test case position.
* added options to toggle OpenDKIM & OpenDMARC
rspamd can provide DKIM signing and DMARC checking itself, so users
should be able to disable OpenDKIM & OpenDMARC. The default is left at
1, so users have to to opt-in when the want to disable the features.
* misc small enhancements
* adjusted start of rspamd
The order of starting redis + rspamd was reversed (now correct) and
rspamd now starts with the correct user.
* adjusted rspamd core configuration
The main configuration was revised. This includes AV configuration as
well as worker/proxy/controller configuration used to control the main
rspamd processes.
The configuration is not tested extensively, but well enough that I am
confident to go forward with it until we declare rspamd support as
stable.
* update & improve the documentation
* add tests
These are some initial tests which test the most basic functionality.
* tests(refactor): Improve consistency and documentation for test helpers (#3012)
* added `ALWAYS_RUN` target `Makefile` recipies (#3013)
This ensures the recipies are always run.
Co-authored-by: georglauterbach <44545919+georglauterbach@users.noreply.github.com>
* adjusted rspamd test to refactored test helper functions
* improve documentation
* apply suggestions from code review (no. 1 by @polarthene)
Co-authored-by: Brennan Kinney <5098581+polarathene@users.noreply.github.com>
* streamline heredoc (EOM -> EOF)
* adjust rspamd test (remove unnecessary run arguments)
Co-authored-by: Brennan Kinney <5098581+polarathene@users.noreply.github.com>
* chore: Co-locate process checking and process restart verification
Extract the test cases for checking a process is running and properly restarts from various test files into a single one:
Core (always running):
opendkim, opendmarc, master (postfix)
ENV dependent:
amavi (amavisd-new), clamd, dovecot, fail2ban-server (fail2ban), fetchmail, postgrey, postsrsd, saslauthd
These now run off a single container with the required ENV and call a common function (the revised version in parallel test cases).
* fix(saslauthd): Quote wrap supervisor config vars
`saslauth.conf` calls `-O` option for most commands defined with an ENV that may be empty/null. This would cause the process to silently fail / die.
This doesn't happen if quote wrapping the ENV, which calls `-O` with an empty string.
Not necessary, but since one of `postgrey` ENV were quote wrapped in `supervisor-app.conf`, I've also done the same there.
* fix(postsrsd): Change supervisor `autorestart` policy to `true`
The PR that introduced the config switched from `true` to `unexpected` without any context. That prevents restart working when the process is killed. Setting to `true` instead will correctly restart the service.
* chore: Remove disabled postgrey test file
`mail_with_postgrey_disabled_by_default.bats` only checked the migrated test cases, removed as no longer serving a purpose.
* tests(refactor): Make `_should_restart_when_killed()` more reliable
The previous version did not ensure that the last checks process was actually restarted, only that it was running.
It turns out that `pkill` is only sending the signal, there can be some delay before the original process is actually killed and restarted.
This can be identified with `pgrep --older <seconds>`. First ensure the process is at a specified age, then after killing check that the process is not running that is at least that old, finally check that there is a younger process actually running.. (_could fail if a process doesn't restart, or there is a delay such as imposed by `sleep` in wrapper scripts for postfix and fail2ban_)
The helper method is not used anywhere else now, move it into this test instead. It has been refactored to accomodate the needs for `--older`, and `--list-full` provides some output that can be matched (similar for `pkill --echo`).
* test(docs): Add inline notes about processes
* chore: Compress test cases into single case with loop
Moves the list of processes into array vars to iterate through instead.
If a failure occurs, the process name is visible along with line number in `_should_restart_when_killed()` to identify what went wrong.
* chore: Handle `FETCHMAIL_PARALLEL=1` process checks as well
* tests: Add test case for disabled ENV
Additional coverage to match what other test files were doing before, ensuring that these ENV can prevent their respective service from running.
* chore: Move `clamd` enabled check to it's own test case
Not sure about this.
It reduces the time of CPU activity (sustained full load on a thread) and increase in memory usage (1GB+ loading signatures database), but as a separate test case it also adds 10 seconds without reducing the time of the test case it was extracted from.
* chore: Make `disabled` variant the 1st test case
* fix: Adjust test cases to pass when using slower wrapper scripts
* tests(refactor): `mail_fetchmail.bats` updated to new format
Additionally merges in the parallel test file.
* chore: Move `config/fetchmail.cf` into separate sub-directory
Keep out of the default base config for tests.
* chore: Change `fetchmail.cf` FQDNs to `.test` TLD
Changed the first configs remote and local user values to more clearly document what their values should represent (_and that they don't need to be a full mail address, that's just what our Dovecot is configured with for login_).
Shifted the `here` to the end of the `is` line. It's optional syntax, only intended to contrast with the remote `there` for readability.
Additionally configured imap protocol. Not tested or verified if that's correct configuration for usage with imap protocol instead. The fetchmail feature tests are currently lacking.
Added an inline doc into the fetchmail test to reference a PR about the importance of the trailing `.` in the config. Updated the partial matching to ensure it matches for that in the value as well.
* chore: Finalize `process-check-restart.bats`
Few minor adjustments. The other ENV for clamd doesn't seem to provide any benefit, trim out the noise. Added a note about why it's been split out.
Fetchmail parallel configs are matching the config file path in the process command that is returned. The `.rc` suffix is just to add further clarity to that.
* fix: Workaround `postconf` write settle logic
After updating `main.cf`, to avoid an enforced delay from reading the config by postfix tools, we can ensure the modified time is at least 2 seconds in the past as a workaround. This should be ok with our usage AFAIK.
Shaves off 2+ seconds roughly off each container startup, reduces roughly 2+ minutes off tests.
* chore: Only modify `mtime` if less than 2 seconds ago
- Slight improvement by avoiding unnecessary writes with a conditional check on the util method.
- Can more comfortably call this during `postfix reload` in the change detection cycle now.
- Identified other tests that'd benefit from this, created a helper method to call instead of copy/paste.
- The `setup email restrict` command also did a modification and reload. Added util method here too.
* tests(fix): `mail_smtponly.bats` should wait for Postfix
- `postfix reload` fails if the service is not ready yet.
- `service postfix reload` and `/etc/init.d/postfix reload` presumably wait until it is ready? (as these work regardless)
* chore: Review feedback - Move reload method into utilities
While working on tests, I noticed that some of the configs being mounted were adding a few seconds to the start-up time of each container. Notably `postfix-*` and `dovecot.conf` config files, which have been extracted out into their own tests with those files moved into a separate config folder.
`tests.bats` has been adapted to the common setup helper, and removed ENV no longer required to run those tests. Future PRs will extract out more tests.
Review may be easier via individual commit diffs and their associated commit messages describing relevant changes.
<details>
<summary>Commit message history for reference</summary>
```markdown
tests(chore): `tests.bats` - Remove redundant config
===
- ONEDIR volume support no longer relevant, this should have been dropped.
- ClamAV ENV no longer relevant as related tests have been extracted already.
- Same with the some of the SpamAssassin ENV config.
- `VIRUSMAILS_DELETE_DELAY` is tested in the file, but doesn't use this ENV at all? (runs a separate instance to test the ENV instead)
- Hostname updated in preparation for migrating to new test helpers. Relevant test lines referencing the hostname have likewise been updated.
```
```markdown
tests(chore): `tests.bats` - Convert to common setup
===
ENV remains the same, but required adding `ENABLE_AMAVIS=1` to bring that back, while the following became redundant as they're now defaulting to explicitly disabled in the helper method:
- `ENABLE_CLAMAV=0`
- `LOG_LEVEL=debug`
- `ENABLE_UPDATE_CHECK=0`
- `--hostname` + `--tty` + standard `--volume` lines
- `-e` option expanded to long-name `--env`, and all `\` dropped as no longer necessary.
`wait_for_finished_setup_in_container` is now redundant thanks to `common_container_setup`.
```
```markdown
tests(refactor): `tests.bats` - Extract out Dovecot Sieve tests
===
Sieve test files relocated into `test/config/dovecot-sieve/` for better isolation.
`dovecot.sieve` was not using the `reject` import, and we should not encourage it? (docs still do):
https://support.tigertech.net/sieve#the-sieve-reject-jmp
```
```markdown
tests: `tests.bats` - Extract out `checking smtp` tests
===
Migrated to the standard template and copied over the original test cases with `_run_in_container` adjustment only.
Identified minimum required ENV along with which mail is required for each test case.
```
```markdown
tests(refactor): `smtp-delivery.bats`
===
- Disabled `ENABLE_SRS=1`, not necessary for these tests.
- Added a SpamAssassin related test (X-SPAM headers) which requires `SA_TAG` to properly pass (or `ENABLE_SRS=1` to deliver into inbox).
- Many lines with double quotes changed to single quote wrapping, and moving out `grep` filters into `assert_output --partial` lines instead.
- Instead of `wc -l` making failures less helpful, switch to the helper method `_should_output_number_of_lines`
- x2 `assert_output` with different EOF style of usage was not actually failing on tests when it should. Changed to assert partial output of each expected line, and count the number of lines instead.
- Added additional comments related to the test cases with a `TODO` note about `SPAMASSASSIN_SPAM_TO_INBOX=1`.
- Revised test case names, including using the common prefix var.
- `tests.bats` no longer needs to send all these emails, no other test cases require them. This affects a test checking a `/mail` folder exists which has been corrected, and a quotas test case adjusted to expect an empty quota size output.
```
```markdown
tests: `tests.bats` - Extract out test cases for config overrides
===
Slight improvement by additionally matching `postconf` output to verify the setting is properly applied.
```
```markdown
tests: `tests.bats` - Extract out Amavis SpamAssassin test case
===
Removes the need for SpamAssassin ENV in `tests.bats`.
```
</details>
- `disabled_clamav_spamassassin`:
- Just shuffling the test order around, and removing the restart test at the end which doesn't make sense.
- `postscreen`:
- Now uses common helper for getting container IP
- Does not appear to need the `NET_ADMIN` capability?
- Reduced startup time for the 2nd container + additional context about it's relevance.
- Test cases are largely the same, but refactored the `nc` alternative that properly waits it's turn. This only needs to run once. Added additional commentary and made into a generic method if needed in other tests.
- `fail2ban`:
- Use the common container IP helper method.
- Postscreen isn't affecting this test, it's not required to do the much slower exchange with the mail server when sending a login failure.
- IP being passed into ENV is no longer necessary.
- `sleep 5` in the related test cases doesn't seem necessary, can better rely on polling with timeout.
- `sleep 10` for `setup.sh` also doesn't appear to be necessary.
- `postgrey`:
- Reduced POSTGREY_DELAY to 3, which shaves a fair amount of wasted time while still verifying the delay works.
- One of the checks in `main.cf` doesn't seem to need to know about the earlier spamhaus portion of the line to work, removed.
- Better test case descriptions.
- Improved log matching via standard method that better documents the expected triplet under test.
- Removed a redundant whitelist file and test that didn't seem to have any relevance. Added a TODO with additional notes about a concern with these tests.
- Reduced test time as 8 second timeouts from `-w 8` don't appear to be required, better to poll with grep instead.
- Replaced `wc -l` commands with a new method to assert expected line count, better enabling assertions on the actual output.
- `undef_spam_subject`:
- Split to two separate test cases, and initialize each container in their case instead of `setup_file()`, allowing for using the default `teardown()` method (and slight benefit if running in parallel).
- `permit_docker`:
- Not a parallel test, but I realized that the repeat helper methods don't necessarily play well with `run` as the command (can cause false positive of what was successful).
- The usual serial to parallel test conversion to utilize the `setup.bash` common setup structure, and adding a `TEST_PREFIX` var for each test case to leverage.
- Standardize on parallel test naming conventions for variables / values.
- More consistent use of `bash -c` instead of `/bin/bash -c` or `/bin/sh -c`.
- Using the `_run_in_container` helper instead of `run docker exec ${CONTAINER_NAME}`.
- Updates tests to use the `check_if_process_is_running` helper.
---
chore: Revise inline docs for the `ssl_letsencrypt` test
- Moves the override to be in closer proximity to the `initial_setup` call, and better communicates the intent to override.
- Removes top comment block that is no longer providing value or correct information to maintainers.
- Revised `acme.json` test case inline doc comments.
