summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Move the default pid file so it doesn't get wiped out by rabbitmqctl reset.bug24095Simon MacMullen2011-08-231-1/+1
|
* Update the man page.Simon MacMullen2011-08-221-6/+14
|
* Merge with defaultRob Harrop2011-08-227-23/+94
|\
| * MNESIA_DIR is not always set. Make sure the directory exists.Simon MacMullen2011-07-281-2/+3
| |
| * ...and get the ownership right.Simon MacMullen2011-07-272-0/+2
| |
| * /var/run/rabbitmq should existSimon MacMullen2011-07-272-0/+2
| |
| * Always write the pid file, and write it in /var/run for debs/RPMs.Simon MacMullen2011-07-272-5/+4
| |
| * It *is* a process. It *has* a pid.Simon MacMullen2011-07-201-3/+3
| |
| * (Untested) attempt at updating OCF script.Simon MacMullen2011-07-201-3/+18
| |
| * Uh, it's probably an idea to write out the pidfile as early as possible, in ↵Simon MacMullen2011-07-191-4/+4
| | | | | | | | case for example prelaunch fails.
| * The problem with kill -0 is that its semantics are "could I send a signal to ↵Simon MacMullen2011-07-151-1/+1
| | | | | | | | this process?" - i.e. it will fail if the process exists but is not owned by the current user. I think this is not quite what we're looking for, so let's use ps -p. I've tested that ps -p does the right thing on Linux and OS X, and it claims to work on FreeBSD and AIX too according to the manpages on the web. That's probably good enough.
| * Do these checks the other way round - if the pid is not running we should ↵Simon MacMullen2011-07-151-7/+7
| | | | | | | | fail, even if the node is up. I think.
| * Be more clear when pidfile read fails.Simon MacMullen2011-07-151-1/+1
| |
| * Show when we've acquired the pid. This would be useful for diagnostics.Simon MacMullen2011-07-151-5/+7
| |
| * We should probably prefix this.Simon MacMullen2011-07-142-3/+3
| |
| * rabbitmqctl wait $PID_FILE.Simon MacMullen2011-07-132-18/+61
| | | | | | | | NB: pid_up/1, with_os/1, system/1 and escape_queues/1 are salvaged from rabbitmq-multi. pid_up/1 used to be is_dead/1 but the new polarity makes more sense.
| * PID_FILESimon MacMullen2011-07-132-2/+9
| |
* | merge bug24354 into default (test_refresh_events/1 sometimes fails)Alexandru Scvortov2011-08-191-1/+5
|\ \
| * | Hypothesis: what's happening is some old channel_created event from a ↵bug24354Simon MacMullen2011-08-191-1/+5
| | | | | | | | | | | | previous test is getting overtaken by the rabbit_tests_event_receiver starting up, and is thus getting through to us. Solution: ignore events with the wrong pid until we get the one we want.
* | | TypoSimon MacMullen2011-08-191-1/+1
| | |
* | | merge bug 24356 into default (pg2_fixed will blow up if compiled with R12B5)Alexandru Scvortov2011-08-192-7/+13
|\ \ \
| * \ \ Merging bug24230 to default, againMatthew Sackman2011-08-1942-803/+1843
| |\ \ \ | | |/ /
| | * | merge bug24194 into defaultMatthias Radestock2011-08-181-4/+7
| | |\ \
| | | * | 'stat' needs to be higher priority toobug24194Matthias Radestock2011-08-181-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Since a typical consumer app will do a queue.declare "assertion" before basic.consume. The down side to this is that a queue.declare can now overtake a 'publish' and thus report incorrect message counts.
| | | * | Allow basic_consume and basic_cancel to match the priority of ack and coMatthew Sackman2011-08-181-4/+6
| | |/ /
| * | | Reverse the polarity, so that it is actually *right*, and make the nearby ↵bug24230Simon MacMullen2011-08-181-3/+6
| | | | | | | | | | | | | | | | comment clearer.
* | | | Remove uses of andalso / orelse where second operand is not a boolean for ↵bug24356Simon MacMullen2011-08-181-10/+22
| |/ / |/| | | | | | | | compatibility with R12.
* | | R12B-3 fix. I don't think this broke anything but it causes the xref CI test ↵Simon MacMullen2011-08-181-1/+1
| | | | | | | | | | | | to fail...
* | | Merging bug24350 to defaultMatthew Sackman2011-08-171-1/+1
|\ \ \
| * | | I refer you to bug24285#c40bug24350Matthew Sackman2011-08-171-1/+1
|/ / /
* | | Merging bug24230 to defaultMatthew Sackman2011-08-1741-798/+1835
|\ \ \ | |/ / |/| |
| * | Merging bug24285 to defaultMatthew Sackman2011-08-176-177/+202
| |\ \
| | * | Cosmeticsbug24285Matthew Sackman2011-08-173-31/+19
| | | |
| | * | All gs:calls should be 'infinity'; cosmetic change to specMatthew Sackman2011-08-171-3/+3
| | | |
| | * | Merging default to bug24285Matthew Sackman2011-08-176-173/+210
| | |\ \
| | | * | refactorAlexandru Scvortov2011-08-153-6/+10
| | | | |
| | | * | start the writer declarativelyAlexandru Scvortov2011-08-121-16/+19
| | | | |
| | | * | start the limiter programaticallyAlexandru Scvortov2011-08-111-13/+8
| | | | | | | | | | | | | | | | | | | | | | | | | Also, how do you name a function that returns a list 'find_child' (singular)?
| | | * | fix specsAlexandru Scvortov2011-08-112-3/+4
| | | | |
| | | * | replace undefined | token() with token()Alexandru Scvortov2011-08-113-15/+13
| | | | |
| | | * | merge from defaultAlexandru Scvortov2011-08-1132-346/+992
| | | |\ \
| | | * | | refactorAlexandru Scvortov2011-08-113-10/+14
| | | | | |
| | | * | | fix dialyzer typesAlexandru Scvortov2011-08-112-4/+6
| | | | | |
| | | * | | catch-all cases, not just undefinedAlexandru Scvortov2011-07-271-1/+1
| | | | | |
| | | * | | cosmeticAlexandru Scvortov2011-07-261-9/+9
| | | | | | | | | | | | | | | | | | | | | | | | Not sure how the misaligned -> happened..
| | | * | | combine function headsAlexandru Scvortov2011-07-261-11/+7
| | | | | |
| | | * | | cosmeticAlexandru Scvortov2011-07-261-4/+4
| | | | | |
| | | * | | cosmeticAlexandru Scvortov2011-07-261-1/+1
| | | | | |
| | | * | | don't needlessly notify queuesAlexandru Scvortov2011-07-261-2/+6
| | | | | |
| | | * | | unbreak amqqueue_process/limiter logicAlexandru Scvortov2011-07-262-8/+10
| | | | | |
View Lorry Controller Status