aboutsummaryrefslogtreecommitdiffstats
path: root/tests
Commit message (Collapse)AuthorAgeFilesLines
* [tests] Correct test package installationJake Hunsaker2021-05-052-10/+9
| | | | | | | | | | | | | | | | First, correct the logs plugin stagetwo test to specify the needed package in a list, and second move the import to within the test class that needs it so that we don't fail on the stageone tests if the package is not present. Second, correct the journal injections that this plugin test does to use `pre_sos_setup()` instead of overriding `setup_mocking()`, and call the former after the mocking has been setup. Finally, add the resulting size of the collected journal if we fail the size requirements of the test. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add test class for when exceptions are expected in an executionJake Hunsaker2021-05-032-6/+89
| | | | | | | | | | | | | | Adds a new `StageOneReportExceptionTest` class that is to be used for tests that should ensure an exception IS hit, for example when testing plugin enablement or invalid options. Included with this commit is a small set of plugin enablement tests. Future tests for component and plugin specific exception handling should be built using this class and design approach. Resolves: #2518 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [Plugin] Make command output size limiting obviousJake Hunsaker2021-04-301-0/+4
| | | | | | | | | | | | | | | | | | | | | When command output is size limited it should be made obvious that it is not the whole output of the command, beyond the collected output (most likely) starting in the middle of a string. To do this, we add a 'truncated' key to the results dict returned by `sos_get_command_output()`, which is set based on if the `AsyncReader` was filled during command execution. From that, first log if a command was truncated. Second, instead of dropping the collected output in the normal plugin directory, instead drop it into `sos_strings` with a `.tailed` extention to mimic how we handled tailed file collections. Finally, add a symlink from the plugin dir with the original filename that points to the `sos_strings` path. Closes: #1332 Resolves: #2506 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Use re.search for assertOutput(Not)ContainsJake Hunsaker2021-04-281-2/+4
| | | | | | | | | | | Standardizes the `assertOutputContains` and `assertOutputNotContains` methods to use `re.search()`, instead of separate parsing methods. `search()` replaces `match()` for efficiency purposes against the collected output. Related: #2507 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [logging] Fix archive debug logging, adjust verbosity levelsJake Hunsaker2021-04-281-1/+20
| | | | | | | | | | | | | | | | | Fixes an issue where archive debug logging was controlled by the use of `--debug` rather than `--verbose`. Removes a superfluous log in `FileCacheArchive.add_link()`. Also, adjusts the different verbosity levels to be more meaninful. Now, the use of `-v` will enable debug logging but will not print those messages to console, `-vv` will print debug logging to console, while `-vvv` will enable archive debug logging which is expected to be significant due to most file operations being logging at some point with archive debug logging. Resolves: #2507 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [process] Collect per-process files from /procJake Hunsaker2021-04-161-0/+29
| | | | | | | | | | | | | | | | | | | | Adds collection of per-process files from /proc/$pid, limited to a number of pids controlled by the new `process.numprocs` option. This option will default to the first 2048 processes. Using a value of `0` for this option will result in all pids being collected. In testing, the initial limit of 2048 process did not significantly raise collection time or archive size. Note however, that collection of 20k or more processes does show a significant increase in plugin execution time and archive size. This is however tempered, by the whole-plugin timeout that sos imposes. Related: #436 Closes: #542 Resolves: #1783 Resolves: #2498 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add a test for default expected plugin enablementJake Hunsaker2021-04-151-1/+56
| | | | | | | | | | | | | | | | Adds a test case to ensure that the plugins we can expect to always be enabled when a "normal" `sos report` command is run are in fact enabled. First, test the distro-independent plugins that do not have specific enablement triggers (and thus should always be enabled). Second, have a distro-specific test for distro-specific plugins that are also expected to always run on those systems. Closes: #2365 Resolves: #2431 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Allow lists for plugin enablement assertionsJake Hunsaker2021-04-151-2/+8
| | | | | | | Updates `assertPlugin(Not)Included` to allow for lists as well as single strings. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add distro specific helper decoratorsJake Hunsaker2021-04-152-3/+14
| | | | | | | | | | | | | | | | | | | Adds helper decorators to define specific test methods for use on specific distributions only. Currently two decorators are available: @redhat_only Only run on fedora, centos, or rhel @ubuntu_only Only run on ubuntu or debian Note that these decorators are only intended for individual `test_*` methods, and will not function to define distro-specific test classes. These should make it easier to write plugin test cases where packaging differences between distributions otherwise makes plugin tests either impossible needlessly complex to write generically. Related: #2431 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Set local distro in StageOne to allow distro-specific testsJake Hunsaker2021-04-152-9/+12
| | | | | | | | | | | | | | Moves the detection of the local system's distribution to StageOne, and defines two constants to reference RH and Ubuntu distribution lists to allow for distro-specific tests. Further, make the expected_warnings_displayed test skipped if the local test system is not a RH-family distro, as those warnings are not expected on non RH-family systems. Related: #2431 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add a pre-setup setup method and smoke testsJake Hunsaker2021-04-152-0/+65
| | | | | | | | | | | | | | | | First, add a "pre-setup setup" method, in the form of `pre_sos_setup()`, that can be used in the way the traditional `setUp()` method would be used (but can't because that's our entry point for executing our sos runs). This method will be executed _prior_ to any mocking. Second, add a smoke test that enables all plugins that exist in the local branch being tested. This will test that doing so does not generate any exceptions and that some expected warnings from select plugins are displayed. Related: #2431 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add plugin tests for basic functionalityJake Hunsaker2021-04-155-0/+147
| | | | | | | | | | | | Adds several plugin tests for plugins that can be expected to run on all distributions, in particular these tests are including mostly Stage Two tests. As such, these tests are only recommended to be run on test machines. Related: #2431 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add a StageTwo timeout testJake Hunsaker2021-04-152-0/+65
| | | | | | Adds a test to ensure our timeout control is working properly. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add StageTwoReportTest as a base test classJake Hunsaker2021-04-151-4/+214
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commits adds a `StageTwoReportTest` from which tests that need to do "light" mocking should be defined. As the name implies, this is the base class for our Stage 2 tests. "Light" mocking for our purposes is currently defined as dropping files in locations on the test system's filesystem, which may include replacing binaries, and installing packages for the durtation of the test case. At the end of the test, packages are uninstalled and mocked files are removed. In the event that a file already exists at a given location on the test system, that file is temporarily renamed with a '.sostesting' extension. On test clean up, once mocked files are removed, any files renamed previously are moved back to their original location/name. Mocking is controlled automatically during test setup, and is based upon test class attributes, similar to how plugin enablement is dependent upon plugin class attributes. For Stage 2 tests cases, this includes: `files` - a list of file names. These names should represent the "real" location on the filesystem where the mocked files will be placed. The mocked content should be placed under a similar file path under tests/test_data E.G. using `files = ['/etc/foo/bar']` will place a copy of `tests/test_data/etc/foo/bar` under `/etc/foo/bar` on the test system. `packages` - a dict of packages to install. The dict keys map to the distribution the package names are used for, e.g. `rhel` or `ubuntu`. The values are lists of package names, optionally with version included. This commit includes just the framework for these tests, and no actual Stage 2 tests are included as of yet. Related: #2431 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [unittests] Update file paths in unittests for new directory layoutJake Hunsaker2021-04-153-15/+15
| | | | | | | | | Updates the filepaths in our unittests for the temp data files to reflect their new location in the updated directory layout. Related: #2431 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Start using avocado for test suiteJake Hunsaker2021-04-1526-272/+715
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commit represents the start of an overhaul of the test suite used by sos. Note that several more commits to follow will be required in order for the test suite to be considered stable. The new test suite will use the avocado-framework to build out new tests. This first part adopts a new 'stageX' naming scheme for our tests as follows: stage0 -> Unittests stage1 -> Basic function tests, no mocking allowed stage2 -> Mocked tests for specific scenarios/regressions stage3 -> Complex setups for layered products/environments At the moment, these unittests are not updated for avocado, though most should still work with `nosetest` directly. A new set of base classes is defined in tests/sos_tests.py which provide the foundation for actual tests cases. This approach entails new test cases subclassing a base class, such as the new `StageOneReportTest`, and setting the `sos_cmd` class attr to the _options_ for an sos report run. By default `sos report --batch` will be run, and targeted to the test job's directory as a tmpdir. Each sos command will be executed once, and all test_* methods within a test case that subclasses `StageOneReportTest` will be checked against the output of that execution. Note that this diverges from avocado's typical approach where each test_* method is performed against a brand new instance of the class (thus meaning any setup including our sos report run would normally be run fresh). However, after speaking with the avocado devel team, this is still seen as a valid pattern for the framework. The current organizational approach is to separate the tests by component rather than stage. For example. `tests/report_tests/` should hold any report-centric tests, and the `plugin_tests` directory therein should be used for plugin-specific tests. As of this commit, there are basic functionality tests under `tests/report_tests/` and a single plugin test under `tests/report_tests/plugin_tests/` to act as a POC. Further, there is a `tests/vendor_tests/` directory for organizing vendor-specific bug/feature tests that are not covered by the generic project-wide tests. A POC test from RHBZ1928628 is available with this commit. Note that in order for these tests to be run properly _without_ installing the current branch to the local system, you will need to run the following command: `PYTHONPATH=tests/ avocado run -t stageone tests/` Related: #2431 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Fix typoEric Desrochers2021-03-231-1/+1
| | | | | | | | | | /var/tmp/sosreport_test/ instead of /var/tmp/sosreport/_test/ Resolves: #2457 Signed-off-by: Eric Desrochers <eric.desrochers@canonical.com> Signed-off-by: Bryan Quigley <code@bryanquigley.com>
* [clean] Fix hostnames not being obfuscated when host plugin wasn't runJake Hunsaker2021-02-171-1/+2
| | | | | | | | | | | | | | | | | | | | | It was discovered that if a report was generated without the host plugin being run, then the hostname parser would not obfuscate the localhost hostname or domain at all. This was due to the fact that while the default_mapping values were being read into the parser's map's dataset correctly, they weren't being loaded into the 'intermediary' dicts that the parser uses to separate short names from domain names. Fix this by reading the contents of the map's dataset dict (which is populated by the map file) into those intermediary dicts that the parser uses. Additionally, safeguard against overloading domain and host obfuscation numbers when loading this way by setting the respective counts to the next-highest number after the map loading is complete. Resolves: #2406 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [policy] switch to sha256 sumBryan Quigley2021-02-091-1/+1
| | | | | | | | | | | | Move cleaner to ignore both sha256/md5s Move collector to use sha256 as well. Move to get hash from stdout Resolves: #2351 Closes: #1204 Signed-off-by: Bryan Quigley <code@bryanquigley.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [sos] Align plugin options with destination namesJake Hunsaker2021-02-032-2/+2
| | | | | | | | | | | | | This commit alters several option long-form names or destination names to align those values in a sensible way. This serves to not only remove some abiguity in option naming in code, but also to make it so that the "effective options" line logged in every sos execution can be direction copy-pasted as a working command. Closes: #2288 Resolves: #2398 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [Policy] Add policy-controlled forbidden pathsJake Hunsaker2021-01-131-0/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This adds policy-controlled forbidden path checking, which should be the final part of implementing "global" forbidden paths. With this commit, policies may now add paths and glob matches for paths which should never be collected in any plugin. Combined with plugin-defined paths and user-defined paths already available, plugins should now be able to be properly restricted from sensitive collections. Note that the way this is implemented is that policies that define the `set_forbidden_paths()` classmethod *extend* this forbidden list as it is built from the subclass(es) that also define one. This way, "top-level" policies do not need to maintain independent copies of entire trees of paths just to add a few specific additional ones that are not forbidden within other policies. This initial commit adds paths that are either very well-known to be ones we should avoid, or are paths that have previously been part of reported issues where these paths/files should not be collected. Closes: #316 Closes: #796 Closes: #919 Closes: #1316 Resolves: #2360 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [PackageManager] Separate PackageManager from policiesJake Hunsaker2021-01-041-1/+2
| | | | | | | | | | | | Moves `PackageManager` out from `sos/policies/__init__.py` into a new `sos/policies/package_managers` subdir. Future commits will aim to canonicalize package manager subclasses for policies to use, and ease the creation of new reusable package managers. Related: #2349 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [policies] Move linux policies into new subdirJake Hunsaker2021-01-042-2/+4
| | | | | | | | | | | Moves the actual policy files for supported distributions out of `sos/policies` and into a new subdir `sos/policies/distros`. Note that `Policy()` still lives in the former, while `LinuxPolicy()` has been moved into `sos/policies/distros/__init__.py` Related: #2349 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add test for newly loaded kernel modulesJake Hunsaker2020-12-111-2/+10
| | | | | | | | | | | | | | | Adds a test to `simple.sh` to check if we loaded any new kernel modules during a normal exeuction. This test will be somewhat limited in the fact that this currently only tests the plugins that are loaded by default on each tested distribution. Also included is a minor typo fix for reporting where IP addresses were not obfuscated, if any. This typo did not affect the actual test however. Resolves: #2326 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Update simple.sh test suiteJake Hunsaker2020-11-191-30/+171
| | | | | | | | | | | | | This patch overhauls tests/simple.sh to provide more granular details on why a test run may have failed, beyond checking for a non-zero exit code or if output was written to stderr. This should also serve as another step towards more easily extendible tests for our automated processes. Resolves: #2305 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [report] Allow users to specify commands and files to skipJake Hunsaker2020-11-032-0/+4
| | | | | | | | | | | | | | | Adds two new options, `--skip-commands` and `--skip-files`, that allow users to selectively skip specific command or file collection instead of having to disable whole plugins to skip those collections. These options are also exposed via `sos collect`, being gated by a version of 4.1 since that is the next scheduled release where we can guarantee this functionality will be present. Closes: #2203 Resolves: #2271 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [general] flake8 first passBryan Quigley2020-08-135-11/+5
| | | | | Signed-off-by: Bryan Quigley <code@bryanquigley.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [travis,style] Upgrade to 20.04 and cover more in testsBryan Quigley2020-08-137-51/+80
| | | | | Signed-off-by: Bryan Quigley <code@bryanquigley.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Update sos.conf locationJake Hunsaker2020-07-211-4/+5
| | | | | | Updates the check for sos.conf with the new location Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [cleaner] Add user-provided keyword obfuscationJake Hunsaker2020-06-171-0/+20
| | | | | | | | | Adds a new parser and map to allow user defined keyword obfuscation. Users may now use the `--keywords` option to have `SoSCleaner` scan lines for matching keywords, and replace them in place like we do for other parsers. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [cleaner] Add hostname parserJake Hunsaker2020-06-171-0/+40
| | | | | | | | | | | | Adds a hostname parser to `sos clean` that will attempt to obfuscate FQDNs matching the hostname of the system that generated the sosreport, as found in sos_commands/host/hostname. Additionally, any domains added via the `--domains` option will also be obfuscated, including any subdomains of the domain(s) specified by the option. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add tests for SoSCleanerJake Hunsaker2020-06-172-0/+98
| | | | | | Adds tests for the mappings and parsers used by SoSCleaner Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [archive] Drop bzip2Bryan Quigley2020-05-261-1/+1
| | | | | | | | | | | | | | | | | | This just leaves xz with gzip as the automatic backup. Bzip2 took the most time by far and only does marginally better than gzip. time 27 Size 10.72MiB --batch -z xz --log-size 1 time 27 Size 16.32MiB --batch -z gzip --log-size 1 time 40 Size 14.17MiB --batch -z bzip2 --log-size 1 time 43 Size 105.21MiB --batch -z xz --all-logs time 64 Size 187.38MiB --batch -z gzip --all-logs time 135 Size 182.86MiB --batch -z bzip2 --all-logs Resolves: #2053 Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Add Timing, summary, and reduce IO for small filesBryan Quigley2020-05-041-9/+20
| | | | | | | | | | | | Timing - see how long each test takes Summary - output the key results at the end Use /dev/shm to store temporary files Make since command actually meaningful Resolves: #2039 Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [plugins] Convert docstrings to class attrsJake Hunsaker2020-04-301-1/+1
| | | | | | | | | | Converts all current plugin docstrings into a `short_desc` attribute, that is now referenced by `Plugin.get_description()`. Closes: #1960 Resolves: #2036 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Fix unittestsJake Hunsaker2020-04-212-3/+3
| | | | | | | | | | Fixes a unittest failure due to an attempt to pass print() a `file` keyword for Red Hat, which is no longer accepted in python3 Additionally fixes 2 string related errors in the test suite since changing from python2 to python3. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [global] Drop use of python six and other 2.X bitsBryan Quigley2020-04-093-15/+6
| | | | | | Resolves: #2006 Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com>
* [travis] Update test script for new binaryJake Hunsaker2020-04-081-9/+9
| | | | | | | Updates the test script `simple.sh` used by Travis to use the new `sos` binary with the `report` subcommand. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [reporting] Move reporting under sos/reportJake Hunsaker2020-04-081-1/+1
| | | | | | Moves the HTML/JSON reporting bits under sos/report. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [global] Fix report import pathsJake Hunsaker2020-04-073-3/+3
| | | | | | | | | | | Updates the uses of 'from sos.plugins' to 'from sos.report.plugins' in order to fix imports across the project with the new tree organization. Additionally, the legacy `sosreport` binary now injects the user's current working directory into the path the python interpreter uses in order to allow local execution from a git checkout. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [Plugin] Add new method for iterating over storage devicesJake Hunsaker2020-04-072-20/+40
| | | | | | | | | | | | | | | | | | | | | | | Adds a new mechanism for iterating over a list of devices with a given command or list of commands. This is to allow us to stop doing ad-hoc device enumeration in multiple plugins and instead provide a consistent behavior for iterating over devices such as disks. Before setting up plugins, sos will now enumerate a list of block devices and fibre devices that is made available to the plugins. From there the new 'add_blockdev_cmd()' method may be passed a command or list of commands that will be iterated over for every device. Commands passed need to include '%(dev)s' in the appropriate place for device name substitution. By default this will iterate over all discovered block devices. The 'devices' parameter may be used to change this, either by passing 'fibre' to instead iterate over fibre devices, or by passing a list of devices to use directly. To filter out or restrict commands to running on specific types of devices, the blacklist and whitelist parameters can be used respectively and can accept either a single regex string or a list of regexes. Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [Policy] Add Container Runtime abstraction classJake Hunsaker2020-04-071-32/+70
| | | | | | | | | | | | | | | | | Adds a `ContainerRuntime()` class to allow policies to specify a container runtime to allow plugins to utilize. The `ContainerRuntime` is intended to allow for the discovery of containers, including specific ones by name, and for the execution of commands inside those containers. This is meant to remove the overhead of manually defining ways to determine an active runtime and if a component is containerized within plugins. Related: #1866 Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] Make simple.sh consistentBryan Quigley2020-03-181-9/+1
| | | | | | | | | | | We no longer have to workaround an odd LXD/Travis interaction so make simple.sh not treat Travis special. Closes: #1921 Resolves: #1983 Signed-off-by: Bryan Quigley <bryan.quigley@canonical.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] add license header to source filesBryn M. Reeves2020-02-269-0/+63
| | | | | | | Resolves: #1958 Signed-off-by: Bryn M. Reeves <bmr@redhat.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] delete random shebangsBryn M. Reeves2020-02-264-8/+0
| | | | | | | | | | | | | Various unit test source files in tests/ have unnecessary shell "shebang" notation specifying a particular Python interpereter. Since these files are executed under the test suite runner there is no need for them to include these lines. Related: #1949 Signed-off-by: Bryn M. Reeves <bmr@redhat.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] remove obsolete test_exe.py scriptBryn M. Reeves2020-02-261-4/+0
| | | | | | | Related: #1949 Signed-off-by: Bryn M. Reeves <bmr@redhat.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] use echo instead of py script in test_{output,shell_out}()Bryn M. Reeves2020-02-261-4/+2
| | | | | | | Related: #1949 Signed-off-by: Bryn M. Reeves <bmr@redhat.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] compare correct string types in test_nonexe_file()Bryn M. Reeves2020-02-261-1/+1
| | | | | | | | | Use a byte string instead of a regular string. Related: #1949 Signed-off-by: Bryn M. Reeves <bmr@redhat.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [tests] look for 'true' in test_exe_file() instead of py scriptBryn M. Reeves2020-02-261-2/+1
| | | | | | | | | | Use the executable 'true' when testing is_executable() on a non-absolute path. Related: #1949 Signed-off-by: Bryn M. Reeves <bmr@redhat.com> Signed-off-by: Jake Hunsaker <jhunsake@redhat.com>
* [policies] fix LinuxPolicy performance regression (test suite)Bryn M. Reeves2020-02-142-21/+21
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Commit 285873a introduces a regression in the performance of LinuxProfile class initialisation: commit 285873a4f753822a88d475a1b030ab622bf4c72e Author: Bryan Quigley <bryan.quigley@canonical.com> Date: Tue Feb 11 15:03:16 2020 -0800 [policies] Detect systemd use instead of hardcoding it All the patch does is to switch from testing self.init to determine whether to use the SystemdInit() class or the basic InitSystem() one, to looking for '/run/systemd/system' in the file system. This has more broad side effects than it might seem since the test suite uses LinuxPolicy() as a mock policy object for a large number of tests. Since SystemdInit() calls out to systemctl to obtain the state of the init system this both increases the run time for the tests and causes high resource use in systemd itself: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1 root 20 0 177196 13504 4512 R 92.7 0.1 5:24.04 systemd This causes the run time for the Python2 and 3 test suite (plus pep8/pycodestyle) to grow from ~5s to over 1m20s on my test system: 285873a~1 real 0m5.683s 285873a real 1m20.353s Allow direct users of LinuxPolicy to override the init system detection by specifying an init= kwarg initialised to an InitSystem-like object directly, and use this in the test suite to avoid the cost of initialising the full SystemdInit() class. Resolves: #1953 Signed-off-by: Bryn M. Reeves <bmr@redhat.com>
generated by cgit (git 2.34.1) at 2024-11-10 07:13:02 +0000