summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Introduce rabbit_vhost:count/0Michael Klishin2019-04-131-1/+5
| | | | (cherry picked from commit dc3c04ffc3862d89d24f55594ae62ab0f8747fdd)
* Fix rabbit_fifo testskjnilsson2019-04-121-1/+2
|
* Fix tests by make quorum queue discovery defensivekjnilsson2019-04-111-1/+7
|
* Merge pull request #1970 from rabbitmq/in-memory-limitsKarl Nilsson2019-04-119-97/+752
|\ | | | | In memory limits for quorum queues
| * correct log effect typekjnilsson2019-04-111-2/+2
| |
| * Correct log effect return typekjnilsson2019-04-112-9/+14
| | | | | | | | | | | | | | It should return a list of effects in recent Ra versions. Also increase ra segments size as this appears beneficial for use with low in-memory limits.
| * Reset in-memory counts after purgingDiana Corbacho2019-04-102-2/+32
| | | | | | | | [#164735591]
| * Fix prefix msgs typesDiana Corbacho2019-04-101-2/+2
| | | | | | | | [#164735591]
| * Rename $ra_no_return to $ra_no_replyDiana Corbacho2019-04-081-1/+1
| | | | | | | | [#164735591]
| * Test that the indexes in the log effect are never lower than previous ↵Diana Corbacho2019-04-041-1/+61
| | | | | | | | | | | | release cursor [#164735591]
| * Apply memory limit on returnsDiana Corbacho2019-04-042-34/+87
| | | | | | | | [#164735591]
| * Merge remote-tracking branch 'origin/master' into in-memory-limitsDiana Corbacho2019-04-0418-244/+850
| |\
| * | Fix in-memory countersDiana Corbacho2019-04-033-63/+198
| | | | | | | | | | | | [#164735591]
| * | Publish in memory statsDiana Corbacho2019-03-281-1/+13
| | | | | | | | | | | | | | | Uses the existing `messages_ram` and `message_bytes_ram` stats [#164735591]
| * | Batch log effectDiana Corbacho2019-03-281-16/+30
| | | | | | | | | | | | [#164735591]
| * | Use in memory limit for quorum queue testsDiana Corbacho2019-03-271-1/+25
| | | | | | | | | | | | | | | All should work as normal, only memory usage it's affected [#164735591]
| * | Return msg in basic.get when in-memory limit is reachedDiana Corbacho2019-03-271-14/+25
| | |
| * | Quorum queue in memory message limitsDiana Corbacho2019-03-276-19/+321
| | | | | | | | | | | | [#164735591]
* | | Merge pull request #1975 from rabbitmq/connection_tracking_no_transactionsLuke Bakken2019-04-111-32/+20
|\ \ \ | | | | | | | | Do not use transactions in connection tracking process.
| * | | Do not use transactions in connection tracking process.Daniil Fedotov2019-04-101-32/+20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The connection tracking process is the only process making changes to the local connections table. There is not much point in having transactions here. Follow-up to #1971 and #1722
* | | | Merge pull request #1976 from rabbitmq/rabbitmq-server-1973-2Michael Klishin2019-04-112-6/+27
|\ \ \ \ | | | | | | | | | | Follow up to #1973 - ensure amq.rabbitmq.log exchange is declared early
| * | | | Add code to declare amq.rabbitmq.log exchange immediately after the / vhost ↵Luke Bakken2019-04-102-5/+26
| | | | | | | | | | | | | | | | | | | | is added or recovered
| * | | | amq.rabbitmq.log really should be an internal exchangeLuke Bakken2019-04-101-1/+1
| | | | | | | | | | | | | | | | | | | | Part of #1973
* | | | | Merge pull request #1978 from rabbitmq/calculate-quorum-memoryKarl Nilsson2019-04-111-1/+1
|\ \ \ \ \ | |/ / / / |/| | | | Update quorum queue supervisor for memory calculation
| * | | | Update quorum queue supervisor for memory calculationDiana Corbacho2019-04-111-1/+1
| | |_|/ | |/| | | | | | | | | | [#161752680]
* | | | Merge pull request #1974 from rabbitmq/rabbitmq-server-1973Luke Bakken2019-04-101-4/+5
|\ \ \ \ | |_|/ / |/| | | Make sure amq.rabbitmq.log is accessible to client connections
| * | | Reduce duplicationMichael Klishin2019-04-101-1/+1
| | | |
| * | | Log amq.rabbitmq.log declaration failures as errorsMichael Klishin2019-04-101-1/+1
| | | |
| * | | Make sure amq.rabbitmq.log is accessible to client connectionsMichael Klishin2019-04-101-3/+4
|/ / / | | | | | | | | | | | | | | | | | | | | | That is, not an internal exchange. Closes #1973, references #1456. [#165243321]
* | | Restore rabbit_connection_tracking_handler:close_connections/3Michael Klishin2019-04-081-0/+6
| | | | | | | | | | | | | | | | | | It is relied on by released 3.7.x versions of rabbitmqctl. References #1971.
* | | Merge pull request #1971 from rabbitmq/connection_tracking_processMichael Klishin2019-04-082-85/+147
|\ \ \ | |/ / |/| | Handle connection tracking events in a separate process.
| * | Handle connection tracking events in a separate process.Daniil Fedotov2019-04-052-85/+147
|/ / | | | | | | | | | | | | | | | | | | | | | | | | Connection tracking event handling may involve DB operations and closing of connections (e.g. for vhost deletion). Because gen_event handles all events in the same process - this may cause message accumulation in the rabbit_event process and memory accumulation. Connection tracking process will only receive connection tracking events and all other events will be dropped, freeing the memory. Follow-up to #1722 [#164353033]
* | Retry Mnesia tables wait in the `quorum_queue` feature flag migrationJean-Sébastien Pédron2019-04-032-3/+3
| | | | | | | | | | | | | | | | | | Before this patch, the migration function waited once 30 seconds then gave up. This was inconsistent with what the boot process does: wait for 30 seconds, 10 times in a row. Now, we do the same thing in the `quorum_queue` migration function, so that it gives up at the same time as the boot process.
* | Merge pull request #1967 from rabbitmq/mk-idempotent-quorum-queue-add-memberMichael Klishin2019-04-022-17/+28
|\ \ | | | | | | Quorum queue: make membership changes [more] idempotent
| * | Quorum queue; make membership changes [more] idempotentMichael Klishin2019-04-012-17/+28
|/ / | | | | | | Per discussion with @kjnilsson.
* | Merge pull request #1961 from rabbitmq/rabbitmq-auth-backend-cache-20D Corbacho2019-03-314-16/+23
|\ \ | | | | | | Cache vhost access requests using client IP address only
| * | Add commentsLuke Bakken2019-03-252-1/+5
| | |
| * | Keep IP address in tuple formatLuke Bakken2019-03-221-7/+2
| | |
| * | Finish converting check_vhost_access/3 to new APILuke Bakken2019-03-224-8/+21
| | |
| * | Modify check_vhost_access/3 to use map of dataLuke Bakken2019-03-222-13/+8
| | | | | | | | | | | | Part of rabbitmq/rabbitmq-auth-backend-cache#20
* | | Merge pull request #1963 from rabbitmq/rabbitmq-server-1962Michael Klishin2019-03-292-0/+20
|\ \ \ | | | | | | | | Follow-up to rabbitmq-server#1962
| * | | Rename function to convert forward slashes, use it in another placeLuke Bakken2019-03-252-6/+17
| | | |
| * | | Merge remote-tracking branch 'gh-1962/master' into rabbitmq-server-1962Luke Bakken2019-03-251-0/+9
| |\ \ \
| | * | | Fix issues with forward slashesClaas Diederichs2019-03-231-0/+9
| | | |/ | | |/| | | | | | | | | Fix issues with forward slashes in RABBITMQ_CONFIG_FILE on windows
* | | | Merge pull request #1960 from ↵Michael Klishin2019-03-292-1/+41
|\ \ \ \ | | | | | | | | | | | | | | | | | | | | rabbitmq/mnesia_use_ram_schema_on_consistency_check Mnesia use ram schema on consistency check
| * \ \ \ Merge branch 'master' into mnesia_use_ram_schema_on_consistency_checkMichael Klishin2019-03-264-165/+546
| |\ \ \ \
| * | | | | Test successfull clustering if there is a garbage schema in the work dirDaniil Fedotov2019-03-221-1/+38
| | | | | |
| * | | | | Use ram schema_location for mnesia consistency checks.Daniil Fedotov2019-03-221-0/+3
| | |/ / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | By default mnesia uses opt_disk schema location, which will use a disk schema if there is a directory. This also applies to the autogenerated directory. In some cases if the data directory has an autogenerated schema directory left from previous versions - the node startup will fail. There is no point in reading a schema on consistency check anyway. [#164807008]
* | | | | Merge pull request #1965 from rabbitmq/quorum-queue-node-removal-cleanupD Corbacho2019-03-284-33/+158
|\ \ \ \ \ | |_|/ / / |/| | | | Purge stale node state from quorum queues
| * | | | Purge stale node state from quorum queueskjnilsson2019-03-274-33/+158
|/ / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When a node is disconnected then removed from the RabbitMQ cluster it is possibly that quorum queues retain some state for this node. This change purges any enqueuer or consumer state for pids relating to nodes that are not in the RabbitMQ cluster. [#164214265]
View Lorry Controller Status