summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* plug leakbug25226Matthias Radestock2012-10-131-11/+9
| | | | | always remove from PendingCh when we receive a msg via gm that we previously received from the channel.
* correct essay & comments regarding the 'delayed confirm' rationaleMatthias Radestock2012-10-122-23/+22
| | | | | | | | | | | Matthew has confirmed that the "we don't know the msg_seq_no until we receive the msg from the channel" reason is bogus. The msg_seq_no is allocated by the channel prior to routing and thus is the same across the master and all slaves. Hence the 'publish' via gm contains all the information we need to issue a confirm. Nevertheless we cannot actually issue the confirm until we've received the message from the channel. The essay now explains the real reason.
* store less in the slave's msg_id_statusMatthias Radestock2012-10-121-20/+16
| | | | | | There is no need to track the ChPid of 'published' and 'confirmed'. Also: neater conversion of msg_id_status on promotion.
* refactor: simplify confirm handling in queueMatthias Radestock2012-10-121-31/+19
| | | | three functions into one
* cosmeticMatthias Radestock2012-10-121-3/+2
|
* re-instate TODOMatthias Radestock2012-10-121-0/+1
|
* a spot of inliningMatthias Radestock2012-10-121-15/+10
| | | | | ...and refactoring; exploiting the fact that maybe_record_confirm_message is a no-op when Confirm == never
* Merge headsSimon MacMullen2012-10-121-2/+4
|\
| * Merge bug25222Simon MacMullen2012-10-120-0/+0
| |\
| * | monitor channel on discardbug25222Matthias Radestock2012-10-121-2/+4
| |/ | | | | | | | | | | just as we do for publish/publish_delivered w/o that there is a 20s delay in slaves cleaning up after channel closure if all the messages sent on the channel were discarded.
* | merge bug25192 into defaultTim Watson2012-10-120-0/+0
|\ \
* \ \ merge default (again)bug25192Tim Watson2012-10-122-18/+11
|\ \ \ | |/ /
| * | Merged bug25195 into defaultEmile Joubert2012-10-126-33/+172
| |\ \ | | |/
| * | Merge defaultbug25195Simon MacMullen2012-10-104-44/+44
| |\ \
| * | | Unused var.Simon MacMullen2012-10-091-1/+0
| | | |
| * | | Remove this master_changed nonsense.Simon MacMullen2012-10-092-13/+0
| | | |
| * | | Ignore process_death, explain why.Simon MacMullen2012-10-091-4/+11
| | | |
* | | | Unused variable.Simon MacMullen2012-10-121-1/+1
| | | |
* | | | Merge in defaultSimon MacMullen2012-10-1210-91/+230
|\ \ \ \ | | |_|/ | |/| |
| * | | Bug 25110 broke ability to invoke "rabbitmqctl stop_app; rabbitmqctl ↵Simon MacMullen2012-10-111-2/+5
| | | | | | | | | | | | | | | | status". Fix that.
| * | | remove a remnant of 'immediate'Matthias Radestock2012-10-111-10/+4
| | | |
| * | | better typesMatthias Radestock2012-10-111-3/+4
| | | |
| * | | merge bug25110 into defaultMatthias Radestock2012-10-112-1/+127
| |\ \ \
| | * | | Remove rabbit_misc:interval_operation/3, it doesn't really work.bug25110Simon MacMullen2012-10-091-13/+0
| | | | |
| | * | | cosmetic: trains stationMatthias Radestock2012-10-081-2/+2
| | | | |
| | * | | correct type signatureMatthias Radestock2012-10-041-2/+1
| | | | |
| | * | | cosmeticMatthias Radestock2012-10-041-11/+15
| | | | |
| | * | | Various tweaks.Simon MacMullen2012-10-022-35/+36
| | | | |
| | * | | Merge defaultSimon MacMullen2012-10-022-1/+136
| | |\ \ \
| | | * | | interval_operation/3Simon MacMullen2012-09-271-2/+15
| | | | | |
| | | * | | Work on older Erlangs.Simon MacMullen2012-09-271-1/+1
| | | | | |
| | | * | | You would think I would at least compile it.Simon MacMullen2012-09-241-1/+1
| | | | | |
| | | * | | Never display negative memory.Simon MacMullen2012-09-241-2/+7
| | | | | |
| | | * | | PluginsSimon MacMullen2012-09-241-1/+21
| | | | | |
| | | * | | Add in memory used by SSL and outgoing connections to the connection / ↵Simon MacMullen2012-09-241-1/+2
| | | | | | | | | | | | | | | | | | | | | | | | channel total.
| | | * | | Ignore dead processes.Simon MacMullen2012-09-241-1/+4
| | | | | |
| | | * | | Move this stuff to its own moduleSimon MacMullen2012-09-242-65/+95
| | | | | |
| | | * | | Go to an entirely supervisor-based way of counting process memory. This has ↵Simon MacMullen2012-09-221-29/+39
| | | | | | | | | | | | | | | | | | | | | | | | the disadvantage that we need to combine channel and connection memory, but we are far more inclusive, accounting for limiters, writers, supervisors etc. In particular this helps a lot when we have thousands of queues since the queue_sup ends up taking a lot of memory in its own right. This seems to bloat rabbit_mgmt_external_stats less (although still some) in the 100k queue case too. Also measure memory use of mgmt_db and msg_store processes and add them appropriately.
| | | * | | Rename this, since it can include memory which does not still belong to a ↵Simon MacMullen2012-09-211-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | process (e.g. after GC).
| | | * | | Account for the mgmt DB as another category, since it can get quite big.Simon MacMullen2012-09-211-1/+3
| | | | | |
| | | * | | Reinstate atomsSimon MacMullen2012-09-191-1/+2
| | | | | |
| | | * | | SimplifySimon MacMullen2012-09-191-3/+1
| | | | | |
| | | * | | As far as the user is concerned, atoms and code are really the same thing ↵Simon MacMullen2012-09-181-2/+1
| | | | | | | | | | | | | | | | | | | | | | | | for memory use.
| | | * | | Move to rabbit.erl, and only invoke erlang:memory/1 once so we do a better ↵Simon MacMullen2012-09-182-51/+55
| | | | | | | | | | | | | | | | | | | | | | | | job of showing results from a single point in time.
| | | * | | ConsistencySimon MacMullen2012-09-181-2/+2
| | | | | |
| | | * | | Try to keep our categories disjoint.Simon MacMullen2012-09-182-8/+12
| | | | | |
| | | * | | rabbit_misc:memory/0.Simon MacMullen2012-09-182-1/+46
| | | | | |
| * | | | | Merge bug25216Simon MacMullen2012-10-101-1/+2
| |\ \ \ \ \ | | | |_|_|/ | | |/| | |
| | * | | | Fix spec broken by bug 25212Simon MacMullen2012-10-101-1/+2
| | | | | |
| * | | | | throttle index walking to the rate at which we can process the resultsbug25216Matthias Radestock2012-10-102-19/+34
| |/ / / / | | | | | | | | | | | | | | | Thus preventing excessive memory use on recovery
View Lorry Controller Status