summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* cosmeticsMatthew Sackman2010-04-281-4/+4
|
* cosmeticsMatthew Sackman2010-04-281-6/+6
|
* Don't use server(Hash) to name the delegate process.Simon MacMullen2010-04-282-3/+2
|
* Explain asyncness.Simon MacMullen2010-04-281-0/+5
|
* Always catch, don't ever allow the server process to die.Simon MacMullen2010-04-281-1/+4
|
* Parallelise communication with multiple nodes.Simon MacMullen2010-04-281-7/+17
|
* Move the gen_server2-via-delegate code into amqqueue; it should only be used ↵Simon MacMullen2010-04-282-40/+32
| | | | there anyway.
* Rename delegate:call and :cast to :invoke and :invoke_async.Simon MacMullen2010-04-284-34/+34
|
* Cache the actual delegate process name, not the result of process_count.Simon MacMullen2010-04-281-15/+14
|
* Remove gen_server2 use, bump up timeouts, tidy GoodRes / BadRes checking.Simon MacMullen2010-04-281-9/+7
|
* cosmeticsMatthew Sackman2010-04-274-26/+18
|
* Merging default into bug 19844Matthew Sackman2010-04-276-126/+99
|\
| * Sanitise boot steps somewhat - a couple of steps had no 'enables' clause, ↵Matthew Sackman2010-04-271-12/+18
| | | | | | | | meaning potentially they could be run last, which would not be desirable (eg exchange recovery after networking up?
| * Merging bug 22651 into defaultMatthew Sackman2010-04-274-111/+77
| |\
| | * explode on attempted double initialisation of queueMatthias Radestock2010-04-271-2/+0
| | | | | | | | | | | | which inidicates a serious bug and thus we shouldn't ignore it
| | * oopsMatthias Radestock2010-04-271-1/+1
| | |
| | * turn queue recovery upside downMatthias Radestock2010-04-274-111/+79
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | instead of the persister pushing recovered messages to the queues, the queues pull recovered messages from the persister. This allows us to perform queue recovery before recording the existince of the queues in mnesia, and thus prevents access to potentially uninitialised queues from other cluster nodes. It also makes the dependency between queues and the persister one way.
| * | Merge headsSimon MacMullen2010-04-271-3/+3
| |\ \ | | |/
| | * cosmeticMatthias Radestock2010-04-261-3/+3
| | |
| * | Merge 22619.Simon MacMullen2010-04-271-0/+1
| |\ \ | | |/ | |/|
| | * Increase the backlog on the listening socket backlog to 128.David Wragg2010-04-161-0/+1
| | | | | | | | | | | | | | | The erlang default is a rather low value of 5. 128 is the default maximum under Linux, so seems a reasonable value to go for.
* | | Fixing up coverage - it would appear the coverage is *very* sensitive about ↵Matthew Sackman2010-04-261-3/+2
| | | | | | | | | | | | how it gets started
* | | Merging default into bug19844Matthew Sackman2010-04-2613-85/+440
|\ \ \ | |/ /
| * | Merging bug 22584 into defaultMatthew Sackman2010-04-262-11/+13
| |\ \
| | * | Add persister_hibernate_after configuration variable.Tony Garnock-Jones2010-04-072-3/+3
| | | |
| | * | Create persister_max_wrap_entries configuration variable.Tony Garnock-Jones2010-04-062-8/+10
| | | |
| * | | merging bug 22647 into defaultMatthew Sackman2010-04-242-8/+6
| |\ \ \
| | * | | change guids to fixed-size binariesMatthias Radestock2010-04-242-8/+6
| | | | | | | | | | | | | | | | | | | | | | | | | This change was cherry-picked from the bug21673 branch and brings 'default' closer to that branch.
| * | | | merging bug 22646 into defaultMatthew Sackman2010-04-243-29/+16
| |\ \ \ \
| | * | | | remove legacy serial seedingMatthias Radestock2010-04-243-29/+16
| | |/ / / | | | | | | | | | | | | | | | thus making the guid generator fully independent of the persister
| * | | | Merging heads of defaultMatthew Sackman2010-04-241-25/+28
| |\ \ \ \
| | * \ \ \ Merging bug 21974 onto defaultMatthew Sackman2010-04-231-25/+28
| | |\ \ \ \ | | | |/ / / | | |/| | |
| | | * | | re-fragment contentMatthias Radestock2010-04-231-25/+28
| | |/ / / | | | | | | | | | | | | | | | so we only produce the minimum number of frames necessary
| * | | | Merge bug22407 into defaultDavid Wragg2010-04-233-3/+3
| |\ \ \ \ | | |/ / / | |/| | |
| | * | | Add build dependencies to Debian / Fedora / MacportsSimon MacMullen2010-03-233-3/+3
| | | | |
| * | | | Merging headsMatthew Sackman2010-04-231-4/+0
| |\ \ \ \
| | * | | | Backed out changeset 1db0ab47b6d5 as the change will create silent failure ↵Matthew Sackman2010-04-231-4/+0
| | | | | | | | | | | | | | | | | | | | | | | | for configurable plugin users who have not moved from the default config location.
| * | | | | Merging bug 22615 onto defaultMatthew Sackman2010-04-223-1/+12
| |\ \ \ \ \
| | * | | | | Cosmetic, and missing specsMatthew Sackman2010-04-221-1/+4
| | | | | | |
| | * | | | | cosmeticMatthew Sackman2010-04-221-1/+0
| | | | | | |
| | * | | | | Roll secondary node coverage into "make start-cover"Simon MacMullen2010-04-212-9/+4
| | | | | | |
| | * | | | | Make start-secondary-cover actually do something, stop Erlang tests blowing up.Simon MacMullen2010-04-202-1/+5
| | | | | | |
| | * | | | | Allow enabling cover for secondary node.Simon MacMullen2010-04-202-0/+10
| | | | | | |
| * | | | | | bug22500 merged into default (OCF/Pacemaker)Marek Majkowski2010-04-213-0/+367
| |\ \ \ \ \ \
| | * | | | | | Only export vars when they're needed, and don't do a full verify unless it's ↵Matthew Sackman2010-04-121-11/+26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | needed - the problem has been that the verify dies because eg the config file doesn't exist, but it doesn't exist because of, say, drbd not being primary+mounted...
| | * | | | | | Merging default into bug 22500Matthew Sackman2010-04-0732-1088/+2409
| | |\ \ \ \ \ \
| | * | | | | | | typeoMatthew Sackman2010-03-161-1/+1
| | | | | | | | |
| | * | | | | | | Cosmetics, and remove bashish -a to generic -e (semantic preserving)Matthew Sackman2010-03-151-6/+6
| | | | | | | | |
| | * | | | | | | Make sure the ocf script goes in /usr/lib/ not /usr/lib64/Matthew Sackman2010-03-151-1/+1
| | | | | | | | |
| | * | | | | | | Minor fixesMatthew Sackman2010-03-111-2/+2
| | | | | | | | |