summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* refactor: make proper use of gb_sets APIMatthias Radestock2013-08-191-5/+5
|
* Use gb_sets rather than sets for performance.Simon MacMullen2013-08-191-6/+6
|
* cosmetics, better var names, and optimise 'DOWN' sendingMatthias Radestock2013-08-171-21/+20
|
* we don't really support demonitor/2, so rip it outMatthias Radestock2013-08-171-11/+8
| | | | | due to the aggregration of all monitors of a single pid into a single monitor, the semantics of demonitor/2 is decidely dodgy
* we only support process monitoring, so enforce that in the APIMatthias Radestock2013-08-171-6/+6
|
* ahemSimon MacMullen2013-08-161-1/+1
|
* Simpler to maintain MonitoredPid -> {Ref, MonitoringPidSet}.Simon MacMullen2013-08-161-66/+32
|
* Explain what this is.Simon MacMullen2013-08-151-0/+3
|
* Store monitors correctly, with a two level dictionary, mapping MonitoredPid ↵Simon MacMullen2013-08-151-10/+57
| | | | -> MonitoringPid -> Ref and MonitoredPid -> Ref -> MonitoringPid.
* Revert b2db7048b839 (which in turn mostly reverted dd08c9204760), i.e. ↵Simon MacMullen2013-08-154-40/+101
| | | | return to the broken bug25644 implementation.
* Stable to defaultSimon MacMullen2013-08-154-101/+40
|\
| * Merged bug25713 into stablerabbitmq_v3_1_5Emile Joubert2013-08-150-0/+0
| |\
| * | Revert dd08c9204760Simon MacMullen2013-08-144-102/+41
| |/
* | Merged bug25572 into defaultEmile Joubert2013-08-136-47/+189
|\ \
| * | Fix inadvertent match of Args. That's what you get for only running the test ↵Simon MacMullen2013-08-131-2/+3
| | | | | | | | | | | | suite.
| * | Man page updatesSimon MacMullen2013-08-091-2/+8
| | |
| * | So it turns out that the policy tests fail, since we now depend on ↵Simon MacMullen2013-08-092-24/+72
| | | | | | | | | | | | rabbit_misc:parse_arguments() to fill in defaults for us. But that's not at all unreasonable I think, and in fact rabbit_tests:control_action() does some rather odd things around options. Rather than rewrite everything, just introduce a new variant that is rather closer to what rabbitmqctl actually does, and use that for policy tests. Oh, and add some more tests, that's always good.
| * | Add support for setting apply-to via rabbitmqctl. Switch to using option for ↵Simon MacMullen2013-08-082-22/+20
| | | | | | | | | | | | priority rather than optional positional argument, with two options now that approach was starting to look confusing.
| * | Gross stupidity.Simon MacMullen2013-08-021-1/+1
| | |
| * | Show apply-to in rabbitmqctl list_policies.Simon MacMullen2013-08-021-1/+1
| | |
| * | "all" is clearer than "both".Simon MacMullen2013-08-022-7/+7
| | |
| * | Reevaluate reified policies on upgrade.Simon MacMullen2013-08-013-3/+47
| | |
| * | Merge defaultSimon MacMullen2013-07-302-5/+50
| |\ \
| | * | Pick an apply-to value heuristically.Simon MacMullen2013-07-301-1/+11
| | | |
| | * | Most of policy apply-to.Simon MacMullen2013-06-242-5/+40
| | | |
* | | | stable to defaultSimon MacMullen2013-08-1329-702/+1070
|\ \ \ \ | |_|_|/ |/| | |
| * | | merge bug25704 into defaultMatthias Radestock2013-08-092-70/+57
| |\ \ \
| | * | | refactor: better name for function applicationMatthias Radestock2013-08-091-6/+4
| | | | | | | | | | | | | | | | | | | | plus lose the type check; we are not that paranoid.
| | * | | tighter specsMatthias Radestock2013-08-081-5/+5
| | | | |
| | * | | Remove delegate beam compatibility check; we don't need that any more.Simon MacMullen2013-08-081-34/+18
| | | | |
| | * | | Use MFA for call and cast, hence avoid sending closures across the network.Simon MacMullen2013-08-081-2/+2
| | | | |
| | * | | Support MFA in delegate:invoke/2 and friends.Simon MacMullen2013-08-081-34/+39
| |/ / /
| * | | Merged stable into defaultEmile Joubert2013-08-071-1/+2
| |\ \ \
| * \ \ \ stable to defaultSimon MacMullen2013-08-0728-632/+1013
| |\ \ \ \
| | * \ \ \ stable to defaultSimon MacMullen2013-08-0628-632/+1013
| | |\ \ \ \
| | | * \ \ \ Merge bug25191Simon MacMullen2013-08-063-56/+96
| | | |\ \ \ \
| | | | * | | | Convert Event set / clear to Alert true / false since that's how it is ↵Simon MacMullen2013-07-301-17/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | everywhere else. Further simplify. Remove a no longer true comment.
| | | | * | | | Inlining and cosmetic.Simon MacMullen2013-07-301-19/+11
| | | | | | | |
| | | | * | | | UntabifySimon MacMullen2013-07-301-21/+21
| | | | | | | |
| | | | * | | | Rename #throttle.conserve_resources to .alarmed_byMichael Klishin2013-07-231-12/+12
| | | | | | | |
| | | | * | | | List connection.blocked in server capabilitiesMichael Klishin2013-07-051-1/+2
| | | | | | | |
| | | | * | | | Deliver all alart notification to handle overlapping alarmsMichael Klishin2013-06-242-24/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | connection.blocked requires us to track resources we are conserving. This means the old logic of determining edge state transitions for alarms does not work any more. Instead of using the old strategy of comparing alarmed node collection sizes, instead of pass around what event the notification is for and simply deliver it to the relevant nodes. This requires that rabbit_alarm event consumers handle duplicate notifications. They already do so after earlier changes on branch bug25191. This makes connection unblocking work correctly in the following sequence of events: * memory alarm set for node A * disk alarm set for node A * memory alarm cleared for node A * disk alarm cleared for node A as well as other similar scenarios with overlapping alarms. This slighly increases internode and intranode message traffic of alarm notifications. Since alarms occur rarely in well-monitored systems, this is a reasonable trade-off.
| | | | * | | | Ignore *.orig filesMichael Klishin2013-06-241-0/+1
| | | | | | | |
| | | | * | | | Use lists:keymember hereMichael Klishin2013-06-211-1/+1
| | | | | | | |
| | | | * | | | Correctly determine duplicate alarms when alarm is clearedMichael Klishin2013-06-211-20/+23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Also, don't bother doing anything on duplicate alarms, per review feedback.
| | | | * | | | Don't log duplicate alarm warningsMichael Klishin2013-06-201-8/+19
| | | | | | | |
| | | | * | | | Make sure the list of alarm resources is uniqueMichael Klishin2013-06-201-1/+8
| | | | | | | |
| | | | * | | | Make sure set_alarm is idempotentMichael Klishin2013-06-191-1/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | By making sure alarms list is unique.
| | | | * | | | Don't crash when an alarm is cleared when there's nothing to clearMichael Klishin2013-06-191-1/+6
| | | | | | | |
| | | | * | | | Merge default into bug25191Michael Klishin2013-06-192-20/+57
| | | | |\ \ \ \
View Lorry Controller Status