summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Merge pull request #2408 from rabbitmq/switch-to-MPL-2.0Michael Klishin2020-07-10275-3737/+1197
|\ | | | | Switch to Mozilla Public License 2.0 (MPL 2.0)
| * Switch to Mozilla Public License 2.0 (MPL 2.0)dcorbacho2020-07-10275-3737/+1197
|/
* Correct a comment typoMichael Klishin2020-07-101-2/+2
|
* Merge pull request #2401 from rabbitmq/redirect-logger-msgs-to-lagerGerhard Lazu2020-07-081-0/+6
|\ | | | | rabbit_prelaunch_early_logging: Configure `rabbit_log` as a Logger handler
| * rabbit_prelaunch_early_logging: Configure `rabbit_log` as a Logger handlerJean-Sébastien Pédron2020-07-061-0/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently, when the new Logger API exposed by Erlang is being used by an application, messages are lost once Lager is configured. That is because Lager does not know about Logger in this context. To fix this, `rabbit_log` is configured as a Logger handler which calls Lager API. This means that Logger messages are passed to Lager. Logger's log level is to set to pass all messages (including debug). Message routing/filtering is done by Lager instead. This change depends on new log/2 function being added to the `rabbit_log` module in rabbitmq-common.
* | Merge pull request #2402 from rabbitmq/metrics-flakinessMichael Klishin2020-07-071-1/+5
|\ \ | |/ |/| Attempt to further de-flake metrics/connection_metric_idemp_test
| * Attempt to further de-flake metrics/connection_metric_idemp_testPhilip Kuryloski2020-07-061-1/+5
|/ | | | | Wait until the connection_metrics report the new connections, before triggering the idempotent action under which they should stay stable.
* Increase timeout in dynamic_ha:nodes_policy_should_pick_master_from_its_paramsPhilip Kuryloski2020-07-061-1/+1
| | | | | | This test still flakes occassionally even with the dynamic wait. The original five second timeout might have been a bit short, as this test takes around 70-90 seconds when successful, so I'm bumping the timeout to 10s to see if it helps.
* One more definition export/import roundtrip caseMichael Klishin2020-07-031-3/+29
|
* Merge pull request #2400 from ↵Michael Klishin2020-07-032-29/+47
|\ | | | | | | | | rabbitmq/mk-resurrect-runtime-parameter-values-as-maps Definition import: pre-format more values as maps where possible
| * Speed up definition import testsMichael Klishin2020-07-031-25/+18
| | | | | | | | | | by avoiding distribution name/port conflicts that do resolve themselves but only after a timeout.
| * rabbit_definitions: format exchange, queue and binding arguments the same ↵Michael Klishin2020-07-032-11/+36
| | | | | | | | | | | | | | way HTTP API would this makes them JSON serialisation-safe without any post-processing of the returned data structure.
| * Revert 460a432ddb2337dad6a7ca350497a8e2aee5063fMichael Klishin2020-07-031-1/+1
|/ | | | | | Definition import should be able to cope with maps. We should move away from using proplists to the extent practically possible, and definition import is one area where we should be able to do that.
* synchronized -> synchronisedPhilip Kuryloski2020-07-021-1/+1
|
* Temporarily switch back an unfortunate atomPhilip Kuryloski2020-07-021-2/+5
| | | | the value was changed in test, but not at its source
* Merge pull request #2399 from ↵Michael Klishin2020-07-012-2/+12
|\ | | | | | | | | rabbitmq/mk-check-if-node-is-mirror-sync-critical-should-ignore-exclusive-queues Exclude exclusive queues from list_local_mirrored_classic_without_synchronised_mirrors
| * list_with_minimum_quorum_for_cli: convert readable names to binariesMichael Klishin2020-07-011-1/+1
| | | | | | | | | | | | so that CLI tools don't have to do any post-processing. References rabbitmq/rabbitmq-cli#432.
| * Exclude exclusive queues from ↵Michael Klishin2020-07-011-1/+11
| | | | | | | | | | | | | | | | list_local_mirrored_classic_without_synchronised_mirrors as they are not actually mirrored. References rabbitmq/rabbitmq-cli#432.
* | Merge pull request #2352 from rabbitmq/honeycombPhilip Kuryloski2020-07-014-714/+4610
|\ \ | |/ |/| Add honeycomb integration for tests
| * Add honeycomb integration for testsPhilip Kuryloski2020-07-014-714/+4610
|/ | | | | | | | We treat the individual test case results as a kind of metric, and collect those in the honeycomb.io observability tool. The intent is to gain a better understanding of the suites from a statistical perspective in order to gain greater insight into issues such as test flakiness.
* Merge pull request #2397 from rabbitmq/rabbitmq-cli-432Michael Klishin2020-07-012-9/+13
|\ | | | | Introduce a way to check if a node is the only cluster member
| * rabbit_nodes:is_single_node_cluster/0: don't check running node countMichael Klishin2020-07-011-1/+1
| | | | | | | | | | | | | | A node with RabbitMQ stopped on it using `rabbitmqctl stop_app` would report no running nodes even in a single node cluster. References #2321
| * Format readable name as a binaryMichael Klishin2020-07-011-1/+1
| | | | | | | | | | | | for CLI tools to output without any post-processing. Part of rabbitmq/rabbitmq-cli#432
| * Introduce a function that checks if a node is the only one in its clusterMichael Klishin2020-07-012-8/+12
|/ | | | Part of rabbitmq/rabbitmq-cli#432
* Merge pull request #2395 from rabbitmq/rabbitmq-server-2394Jean-Sébastien Pédron2020-06-307-1/+35
|\ | | | | Restore CTL_ERL_ARGS env variables
| * Re-enable CTL_ERL_ARGS and dist port settingsLuke Bakken2020-06-307-1/+35
|/ | | | Fixes #2394
* Clarify that the delay is in ms in these log messagesMichael Klishin2020-06-271-2/+2
| | | | References #1627, #2276.
* Avoid using unfortunate terms in more placesMichael Klishin2020-06-268-113/+113
| | | | | | | | We have switched all doc guides to use "follower", "mirror" or "secondary" years ago but some comments and test suites were never updated, even though their changes would not be breaking. This transition is a medium-term WIP.
* Avoid using unfortunate terms in more placesMichael Klishin2020-06-265-12/+12
| | | | | | | | | We have switched all doc guides to use "mirror" or "secondary" years ago but these were never updated. Renaming functions and record/HTTP API fields (including CLI tools) would be major a breaking change, so they will be aliased or renamed with a lot more extensive review in the future.
* Avoid using unfortunate terms in more placesMichael Klishin2020-06-2611-82/+82
| | | | | | | | | We have switched all doc guides to use "mirror" or "secondary" years ago but these were never updated. Renaming functions and record/HTTP API fields (including CLI tools) would be major a breaking change, so they will be aliased or renamed with a lot more extensive review in the future.
* Avoid using unfortunate terms in code comments and log messagesMichael Klishin2020-06-2613-85/+85
| | | | | | | | | We have switched all doc guides to use "mirror" or "secondary replica" years ago but these were never updated. Renaming functions and record/HTTP API fields (including CLI tools) would be major a breaking change, so they will be aliased or renamed with a lot more extensive review in the future.
* Merge pull request #2396 from rabbitmq/rabbitmq-server-2384Michael Klishin2020-06-262-21/+32
|\ | | | | Import definitions in the post-launch phase
| * Avoid seeding default data if there are configured definitions to load on bootMichael Klishin2020-06-262-9/+27
| | | | | | | | | | | | | | | | | | This is a behavior we use to provide because definitions were imported before the empty DB boot step. Now when the definitions are imported in the post-launch phase, we'd always end up with default data seeded in addition to definitions being imported. References #2384
| * Export definitions on boot after all plugins have been enabledMichael Klishin2020-06-261-12/+5
|/ | | | | | | This way definitions that depend on a plugin, such as federation upstreams or exchanges of custom types, can be imported successfully. Closes #2384
* Merge pull request #2393 from rabbitmq/upgrade-cuttlefishJean-Sébastien Pédron2020-06-251-1/+1
|\ | | | | Makefile: Upgrade Cuttlefish from 2.2.0 to 2.3.0
| * Makefile: Upgrade Cuttlefish from 2.2.0 to 2.3.0Jean-Sébastien Pédron2020-06-251-1/+1
| | | | | | | | | | This fixes a crash of Cuttlefish when a line is made only of whitespaces.
* | Merge pull request #2392 from fenollp/nohipeMichael Klishin2020-06-257-253/+3
|\ \ | |/ |/| drop deprecated HiPE compilation
| * Keep the flag for backwards config compatibility.Pierre Fenoll2020-06-241-0/+9
| | | | | | | | Signed-off-by: Pierre Fenoll <pierrefenoll@gmail.com>
| * drop deprecated HiPE compilationPierre Fenoll2020-06-247-260/+1
| | | | | | | | Signed-off-by: Pierre Fenoll <pierrefenoll@gmail.com>
* | Fix a typo spotted by @VVDMichael Klishin2020-06-241-1/+1
|/ | | References #2389, #2387
* Merge pull request #2389 from ↵Michael Klishin2020-06-241-0/+28
|\ | | | | | | | | rabbitmq/restore-dist-port-configuration-for-windows-service scripts/rabbitmq-service.bat: Configure Erlang distribution port
| * scripts/rabbitmq-service.bat: Configure Erlang distribution portJean-Sébastien Pédron2020-06-191-0/+28
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | As said in the comment, this is normally done at runtime now (see rabbitmq/rabbitmq-server#2180). However, we must configure the distribution on the command line specifically for the Windows service. Until this patch, the distribution was enabled with the correct nodename, but the TCP port configuration was not set. This caused a regression because the selected TCP port was random instead of the expected default of 25672. This patch restores the previous behavior. Fixes rabbitmq/rabbitmq-server#2387.
* | Update erlang.mkJean-Sébastien Pédron2020-06-231-3/+71
|/
* Merge pull request #2378 from rabbitmq/load-conf-before-erlang-dist-setupJean-Sébastien Pédron2020-06-101-4/+4
|\ | | | | rabbit_prelaunch: Load configuration before Erlang dist. is started
| * rabbit_prelaunch: Load configuration before Erlang dist. is startedJean-Sébastien Pédron2020-06-101-4/+4
|/ | | | | | | | The reason is that kernel's `net_ticktime` can be configured in the Cuttlefish or Erlang configuration files. If we start the distribution first, then the configuration parameter is ignored. Fixes #2377.
* Merge pull request #2375 from rabbitmq/rabbit-fifo-int-test-stabilityPhilip Kuryloski2020-06-101-10/+14
|\ | | | | Attempt to increase rabbit_fifo_int_SUITE stability
| * Attempt to increase rabbit_fifo_int_SUITE stabilityPhilip Kuryloski2020-06-091-10/+14
| | | | | | | | | | Wait for messages to have been enqueued before proceeding with the body of the test case
* | Merge pull request #2374 from rabbitmq/metrics-suite-test-stabilityPhilip Kuryloski2020-06-101-3/+7
|\ \ | | | | | | Make connection_metric_idemp_test example rather than property based
| * | Replace connection_metric_idemp_test with specific casesPhilip Kuryloski2020-06-091-3/+7
| |/ | | | | | | | | | | | | | | rather than a property based test. The property test flakes occassionally, and without consistent input values. Idempotency should be properly demonstrated with 2 or more attempts at the action anyway.
* | Merge pull request #2376 from rabbitmq/simple-ha-test-stabilityPhilip Kuryloski2020-06-101-0/+16
|\ \ | | | | | | Stricter setup in the confirms_survive_* tests
View Lorry Controller Status