summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Demonitor on channel_cleanupMatthew Sackman2011-02-111-7/+8
|
* Log in the case of an uncontrolled, unexpected channel exitMatthew Sackman2011-02-111-1/+4
|
* Remove needless variableMatthew Sackman2011-02-111-1/+1
|
* Correct rabbit test for new channel exit sequenceMatthew Sackman2011-02-111-1/+1
|
* Always erase both channel and ch_pid. This is lovely - it makes it a proper ↵Matthew Sackman2011-02-112-6/+5
| | | | bidirectional mapping. All java tests now pass again.
* drop CPS in favour of explicit APIMatthew Sackman2011-02-112-7/+10
|
* Move the state modification into rollback_and_notifyMatthew Sackman2011-02-111-15/+14
|
* Remove boolean shouldClose from response of bin_gen:map_exceptionMatthew Sackman2011-02-113-17/+13
|
* cosmetic - moving protocol field up.Matthew Sackman2011-02-111-7/+6
|
* Remove hardcoded protocol now channels have the real protocolMatthew Sackman2011-02-111-5/+4
|
* Remove annoying documentationMatthew Sackman2011-02-111-86/+0
|
* WhoopsMatthew Sackman2011-02-101-1/+1
|
* The channel.close_ok from the client might be immediately followed by a ↵Matthew Sackman2011-02-101-0/+3
| | | | channel.open. As such we must make sure that we send the channel.open to a new channel
* close_channel is no longer needed in reader, which means that closing as a ↵Matthew Sackman2011-02-101-23/+0
| | | | channel state is no longer needed.
* Always log the error from the channel; can now guarantee exceptions that ↵Matthew Sackman2011-02-102-18/+9
| | | | turn up in the reader are hard exceptions
* Log soft channel errorsMatthew Sackman2011-02-101-0/+2
|
* cosmeticMatthew Sackman2011-02-101-2/+2
|
* Tidy channel termination (state=closing implies rollback_and_notify done). ↵Matthew Sackman2011-02-102-22/+19
| | | | Erase both ch_pid and channel from reader state on DOWN (required for soft errors in which the channel issues the close, waits for the close_ok, and then stops normally)
* If we receive a channel.close whilst we're closing the channel already ↵Matthew Sackman2011-02-101-0/+3
| | | | (presumably client and server initiated close at the same time), then make sure we still send back the close_ok
* server initiated channel.close: keep channel alive until ch.close_ok is ↵Matthew Sackman2011-02-101-3/+22
| | | | received. Make channel responsible for deciding on the severity of the error
* Ignore methods to the channel if we're closingMatthew Sackman2011-02-101-0/+3
|
* Make client-initiated channel.close work in new schemeMatthew Sackman2011-02-103-17/+21
|
* Pass protocol through to channelMatthew Sackman2011-02-102-15/+19
|
* Merged bug23811 (no need for amqqueue:notify_sent to use delegate)Simon MacMullen2011-02-101-1/+1
|\
| * use a plain cast, not delegateMatthew Sackman2011-02-071-1/+1
| |
* | Merge bug23813 (abstract rabbit out of delegate)Simon MacMullen2011-02-1017-52/+759
|\ \
| * \ merge bug23806 into default (Re-declaring a queue on a different cluster ↵Alexandru Scvortov2011-02-094-18/+57
| |\ \ | | | | | | | | | | | | node fails)
| | * | Added test which covers the mysteriously dead queue on queue declaration, ↵Matthew Sackman2011-02-081-2/+30
| | | | | | | | | | | | | | | | and hence misc:is_process_alive
| | * | I need a much longer timeout for these tests...Matthew Sackman2011-02-081-1/+1
| | | |
| | * | Merging default to bug23806Matthew Sackman2011-02-086-20/+100
| | |\ \
| | * | | abstract out is_process_aliveMatthew Sackman2011-02-083-19/+29
| | | | |
| | * | | remove non-rpc'd call to is_process_alive - now matches the rpc version in ↵Matthew Sackman2011-02-071-4/+5
| | | | | | | | | | | | | | | | | | | | amqqueue_process
| * | | | Merge bug23792 into defaultSimon MacMullen2011-02-095-15/+13
| |\ \ \ \ | | |_|/ / | |/| | |
| | * | | Make tcp listen options configurableMatthew Sackman2011-02-025-15/+13
| | | | |
| * | | | merge bug23814 into default (misc:with_exit_handler is missing a match for ↵Alexandru Scvortov2011-02-081-2/+5
| |\ \ \ \ | | | | | | | | | | | | | | | | | | nodedown and shutdown)
| | * | | | Make with_exit_handler more robust and able to cope with the other forms of ↵Matthew Sackman2011-02-071-2/+5
| | | |/ / | | |/| | | | | | | | | | | | shutdown and nodedown
| * | | | Merging bug22698 to defaultMatthew Sackman2011-02-084-15/+88
| |\ \ \ \
| | * | | | remove listeners when the rabbit application is stoppedAlexandru Scvortov2011-02-071-5/+3
| | | | | |
| | * | | | Refactoring, cosmetics, consistency, return codesMatthew Sackman2011-02-072-14/+11
| | | | | |
| | * | | | merge from defaultAlexandru Scvortov2011-02-035-10/+30
| | |\ \ \ \
| | * | | | | add testAlexandru Scvortov2011-02-033-6/+39
| | | | | | |
| | * | | | | make cluster nofitication a boot step; don't print to screenAlexandru Scvortov2011-01-312-13/+33
| | | | | | |
| | * | | | | cleanup after rabbit goes down on a clustered nodeAlexandru Scvortov2011-01-311-6/+12
| | | | | | |
| | * | | | | monitor rabbits on clustered nodesAlexandru Scvortov2011-01-312-6/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When a rabbit starts, it, * informs every other running node in the cluster that it has started, and * registers every other running node in the cluster as running. This should handle every possible case of nodes joining clusters, since every join should be followed by that node starting. We can't just monitor 'rabbit' on node ups, since those come in when mnesia joins the cluster. The rabbit process starts up a bit later.
| | * | | | | merge from defaultAlexandru Scvortov2011-01-312-0/+6
| | |\ \ \ \ \
| | | * \ \ \ \ merge from default (rabbitmq-multi works again)Alexandru Scvortov2011-01-253-2/+4
| | | |\ \ \ \ \
| | | * | | | | | monitor the rabbit app on some remote nodesAlexandru Scvortov2011-01-251-0/+5
| | | | | | | | |
| | | * | | | | | register the almost-top rabbit process as 'rabbit'Alexandru Scvortov2011-01-251-0/+1
| | | | | | | | |
| * | | | | | | | merge bug23804 into default (ssl connections dying prior to R14)Alexandru Scvortov2011-02-071-3/+7
| |\ \ \ \ \ \ \ \ | | |_|_|_|_|/ / / | |/| | | | | | |
| | * | | | | | | Keep the same formate as ssl:connection_infoSimon MacMullen2011-02-071-6/+6
| | | | | | | | |
View Lorry Controller Status