summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Update rabbitmq-plugins usageTim Watson2014-03-141-2/+16
|
* Improve synchronisation between plugins file and running brokerTim Watson2014-03-141-22/+64
| | | | | | | | | | | Since multiple brokers (on a host) may share the enabled-plugins file, we may get out of sync. Diff any changes stemming from enable/disable operations against the plugins actively running on the broker. Extend the 'list' operation to display plugins that are [A] active but not explicitly (or implicitly) enabled in the file and [I], those that are explicitly (or implicitly) enabled but not actively running in the broker.
* Merge default into bug24926Tim Watson2014-03-137-64/+251
|\
| * Refactor: Simplify boot/cleanup step handling and unify modulesTim Watson2014-03-133-175/+91
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Isolate boot steps to a set of applications and avoid using ets to track which have run. This also simplifies cleanup step execution. The force_reload code isn't needed, since the real issue preventing updated versions of modules from being found lay in the proper expansion of (the correct set of) .ez archives, which was fixed in revision 1918e77. Unify all the boot handling code under rabbit.erl once again and now completely remove rabbit_boot, since it's no longer needed. Remove unused exports and tidy.
| * We missed another change when moving code back into rabbit.erlTim Watson2014-03-051-1/+1
| |
| * Fix an oops in rabbit:stop/0 and apply consistent parameter names in r_pluginsTim Watson2014-03-042-4/+4
| |
| * Inline some more and get closer still to defaultTim Watson2014-03-043-24/+26
| |
| * Reduce the distance to default a bit moreTim Watson2014-03-043-47/+38
| |
| * Publish plugin changes via rabbit_eventTim Watson2014-03-032-3/+31
| |
| * Fix plugin expansion during runtime configuration changesTim Watson2014-02-281-14/+17
| | | | | | | | | | | | | | Blowing away the plugin expand directory when we make runtime changes is wrong and we don't need to strip away expanded plugins when they're disabled, since the directory is cleaned on each restart and leaving the files in place isn't an issue.
| * Simply application:get_env calls and reduce distance to default againTim Watson2014-02-213-11/+9
| |
| * Merge default into bug24926Tim Watson2014-02-2110-107/+356
| |\
| | * Cosmetic - reduce the distance to default a bit moreTim Watson2014-02-212-33/+30
| | |
| | * Fix various oopses and reduce the distance to default some moreTim Watson2014-02-182-24/+3
| | |
| | * Try and reduce the distance to defaultTim Watson2014-02-182-193/+207
| | |
| | * Add a type spec for prepare_boot_tableTim Watson2014-02-181-8/+9
| | |
| | * Simplifying refactorTim Watson2014-02-182-76/+11
| | | | | | | | | | | | | | | | | | | | | | | | Since the "rabbit" process /always/ starts before everything else, handle boot table creation at that point and avoid the ets ownership issue and corresponding messiness. The previous worries about handling situations where we want to run rabbit_boot code whilst rabbit is offline aren't an issue for us.
| | * Use a gen_server to keep the boot steps table aroundTim Watson2014-02-182-81/+83
| | |
| | * Merge default into bug24926Tim Watson2014-02-1810-218/+539
| | |\
| | | * Merge default into bug24926Tim Watson2014-02-178-65/+98
| | | |\
| | | * | Avoid verbosityTim Watson2014-02-171-9/+6
| | | | |
| | | * | Don't bother printing out reload errorsTim Watson2014-02-171-1/+1
| | | | |
| | | * | SimplifyTim Watson2014-02-171-3/+2
| | | | |
| | | * | If you're going to catch, actually handle failuresTim Watson2014-02-171-1/+1
| | | | |
| | | * | Cosmetic (ish)Tim Watson2014-02-171-4/+1
| | | | |
| | | * | merge default into bug24926Tim Watson2014-02-1110-218/+546
| | | |\ \
| | | | * | Refactor - remove unused functions/comments/todosTim Watson2014-01-125-40/+14
| | | | | |
| | | | * | Merge default into bug24926Tim Watson2014-01-1219-82/+137
| | | | |\ \
| | | | * | | Refactor (simplify application stop handling)Tim Watson2014-01-122-50/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The offline plugin handling does a fine job of detecting which apps need to be stopped/unloaded when disabling a plugin at runtime. We previously had problems accidentally stopping things like stdlib and mnesia by accident, but only because we went searching for the app dependencies ourselves! Instead, we stop only those applications (i.e., plugins) that rabbit_plugins_main has requested. In doing so, we remove a considerable amount of (new) digraph hacking from app_utils and reduce the complexity of rabbit_boot somewhat.
| | | | * | | Handle hard vs. soft boot step dependenciesTim Watson2014-01-101-15/+37
| | | | | | |
| | | | * | | Merge default into bug24926Tim Watson2014-01-1030-971/+1451
| | | | |\ \ \ | | | | | | | | | | | | | | | | | | | | | | | | Merge conflicts in rabbit.erl, pertaining to app startup procedure.
| | | | * | | | Tidy/Refactor boot step handlingTim Watson2013-12-161-14/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Simplify the boot step loading/application code and switch from using a discrete rabbit_cleanup_step module attribute; We now search the regular boot steps directly for ‘cleanup’ entries.
| | | | * | | | refactor (renaming)Tim Watson2013-11-262-4/+4
| | | | | | | |
| | | | * | | | De-duplicate dependency/graph building functionsTim Watson2013-11-262-26/+40
| | | | | | | |
| | | | * | | | Remove lots of io:format debugging and tidy up rpc error messagesTim Watson2013-11-262-14/+12
| | | | | | | |
| | | | * | | | Re-work boot step handlingTim Watson2013-11-213-46/+146
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We need to support interleaving plugin boot steps with rabbit’s own (internal) ones during the boot/start sequence, and also running them independently during runtime activation. This commit also drops the use of an ets table to track which cleanup steps we’ve run, and simplifies the boot step execution procedures.
| | | | * | | | Refactor (de-duplicate) module attribute searchTim Watson2013-11-201-16/+17
| | | | | | | |
| | | | * | | | Various boot procedure improvementsTim Watson2013-11-204-37/+91
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We ensure that all dependent applications are stopped, but also that dependencies shared by multiple applications (e.g., amqp_client) are not.
| | | | * | | | merge default into bug24926Tim Watson2013-11-187-56/+113
| | | | |\ \ \ \
| | | | * | | | | Rework boot step handlingTim Watson2013-11-182-8/+14
| | | | | | | | |
| | | | * | | | | handle cleanup steps properlyTim Watson2013-11-061-13/+18
| | | | | | | | |
| | | | * | | | | track boot status throughout a node's lifetimeTim Watson2013-11-062-55/+111
| | | | | | | | |
| | | | * | | | | Allow offline plugin re-configurationTim Watson2013-11-051-1/+4
| | | | | | | | |
| | | | * | | | | Ensure we pass the nodename on windowsTim Watson2013-11-051-1/+5
| | | | | | | | |
| | | | * | | | | Allow for runtime disabling of pluginsTim Watson2013-11-054-17/+53
| | | | | | | | |
| | | | * | | | | Restructure our boot procedure for more flexible boot-step handlingTim Watson2013-11-056-147/+245
| | | | | | | | |
* | | | | | | | | stable to defaultSimon MacMullen2014-03-111-3/+4
|\ \ \ \ \ \ \ \ \
| * \ \ \ \ \ \ \ \ Merge bug 26063Simon MacMullen2014-03-110-0/+0
| |\ \ \ \ \ \ \ \ \
| * | | | | | | | | | Don't corrupt the node montor state even if we are about to shut down.Simon MacMullen2014-03-111-3/+4
| |/ / / / / / / / /
* | | | | | | | | | Be a tiny bit more correctSimon MacMullen2014-03-101-1/+1
| | | | | | | | | |
View Lorry Controller Status