Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Sleep briefly, and explain why. | Simon MacMullen | 2014-10-20 | 1 | -0/+8 |
| | |||||
* | Rename GM:terminate/2 to GM:handle_terminate/2 so that it doesn't conflict ↵ | Simon MacMullen | 2014-09-17 | 1 | -4/+4 |
| | | | | with the same callback in gen_server. | ||||
* | Improve self-documentingness.bug26286 | Simon MacMullen | 2014-07-15 | 1 | -2/+11 |
| | |||||
* | Revert bug26171, and ignore messages before the initial catchup. | Simon MacMullen | 2014-07-14 | 1 | -31/+46 |
| | |||||
* | Remove some unreachable code - members_state will be filled once we have ↵bug26171 | Simon MacMullen | 2014-06-27 | 1 | -17/+0 |
| | | | | joined, and now we join immediately at startup. | ||||
* | Merge bug24929. | Simon MacMullen | 2014-06-23 | 1 | -27/+30 |
|\ | |||||
| * | Merge in bug24929 | Simon MacMullen | 2014-06-02 | 1 | -27/+30 |
| |\ | |||||
| | * | Abstract the initial catchup into its own function. Make sure we call ↵ | Simon MacMullen | 2014-05-12 | 1 | -21/+20 |
| | | | | | | | | | | | | check_neighbours beforehand, since that sets left and right. | ||||
| | * | Short-circuit the initial catchup message from the left, so that we are sure ↵ | Simon MacMullen | 2014-05-12 | 1 | -15/+19 |
| | | | | | | | | | | | | we receive it before we consider ourselves joined. | ||||
* | | | Add instrumenting for monitors and Mnesia transactions. It actually all ↵ | Simon MacMullen | 2014-06-23 | 1 | -2/+14 |
|/ / | | | | | | | seems to work now. | ||||
* | | Merge in stable | Simon MacMullen | 2014-06-02 | 1 | -7/+8 |
|\ \ | |||||
| * | | Small refactor | Simon MacMullen | 2014-06-02 | 1 | -7/+7 |
| | | | |||||
| * | | Silence warning | Simon MacMullen | 2014-06-02 | 1 | -1/+1 |
| | | | |||||
* | | | Don't be so intrusive into gm.erl. | Simon MacMullen | 2014-06-02 | 1 | -59/+37 |
| | | | |||||
* | | | Instrument certain calls to gs2:call/3 and gs2:cast/2, and allow PropEr to ↵ | Simon MacMullen | 2014-05-29 | 1 | -38/+59 |
|/ / | | | | | | | determine when they proceed. That makes things a lot more deterministic (although not fully) and lets us reproduce bug 26171 (small whoop). | ||||
* | | Don't prioritise DOWNs from the left. Also explain why we prioritise in ↵bug26210 | Simon MacMullen | 2014-05-29 | 1 | -1/+14 |
|/ | | | | various ways. | ||||
* | Fix broken invocation.bug26173 | Simon MacMullen | 2014-05-12 | 1 | -1/+2 |
| | |||||
* | blank members_state before maybe_erase_aliasesbug26084 | Matthias Radestock | 2014-04-23 | 1 | -7/+6 |
| | | | | | | | | | | thus reverting b01e5808866d We are the only member left. So our usual alias erasure condition of "acks = pubs" needs to go since otherwise erasure could be deferred indefintely, waiting for acks to arrive that never will. Blanking the members_state effectively sets acks = pubs for all aliases, thus enabling their unconditional erasure. | ||||
* | don't erase aliases quite so eagerly on DOWNbug26119 | Matthias Radestock | 2014-04-23 | 1 | -8/+9 |
| | | | | | since that can result in completely forgetting about a member when there are still acks making their way around the right. | ||||
* | maintain gm ?GROUP_TABLE abstraction | Matthias Radestock | 2014-04-22 | 1 | -3/+3 |
| | | | | in yet another place | ||||
* | refactor: extract gm view change handling | Matthias Radestock | 2014-04-21 | 1 | -29/+18 |
| | |||||
* | cosmetic: consistency | Matthias Radestock | 2014-04-21 | 1 | -11/+12 |
| | |||||
* | make gm record_new_member_in_group tx side effect free | Matthias Radestock | 2014-04-21 | 1 | -22/+14 |
| | |||||
* | refactor: tweak and push through gm read_group abstraction | Matthias Radestock | 2014-04-21 | 1 | -10/+13 |
| | |||||
* | simplify: get rid of tmp var | Matthias Radestock | 2014-04-21 | 1 | -57/+50 |
| | |||||
* | enforce gm GROUP_TABLE abstraction | Matthias Radestock | 2014-04-21 | 1 | -21/+17 |
| | | | | | | ...by using mnesia:write/3 instead of /1 and generally abstract mnesia reading/writing and make it dirty reads obvious. | ||||
* | blank members_state after maybe_erase_aliases | Matthias Radestock | 2014-04-21 | 1 | -7/+9 |
| | | | | since the latter operates on the former | ||||
* | refactor: simplify gm:internal_broadcast | Matthias Radestock | 2014-04-21 | 1 | -19/+12 |
| | | | | eliminate branching | ||||
* | go back to from members_changed/4 to /3 | Matthias Radestock | 2014-04-17 | 1 | -7/+5 |
| | |||||
* | less lazy death notifications | Matthias Radestock | 2014-04-13 | 1 | -21/+9 |
| | | | | | | | | | | | When we receive a DOWN, it's possible that there are no messages from the dead member floating around anyway. So rather than waiting for some other activity to occur first, it is perfectly safe enough to always call maybe_erase_aliases. That way we get slightly less lazy death notification in this case (i.e. we can then be certain that either the dead member is fully removed immediately or there must still be messages from the dead member floating around. No other possibility exists). | ||||
* | ensure callback_view_changed is called whenever the view may have changed | Matthias Radestock | 2014-04-13 | 1 | -8/+8 |
| | | | | | | We had missed a case here. Ditto for check_neighbours. | ||||
* | Update copyright for 2014bug25940 | Simon MacMullen | 2014-03-17 | 1 | -1/+1 |
| | |||||
* | stable to default | Simon MacMullen | 2014-01-17 | 1 | -36/+70 |
|\ | |||||
| * | Avoid "Absurdly large distribution output data buffer" death. | Simon MacMullen | 2014-01-16 | 1 | -36/+70 |
| | | |||||
* | | Revert changes to add proc_name via gen_server2, instead do it explicitly ↵bug25910 | Simon MacMullen | 2014-01-10 | 1 | -2/+2 |
| | | | | | | | | with a helper macro. | ||||
* | | Merge in default | Simon MacMullen | 2014-01-10 | 1 | -1/+2 |
|\ \ | |/ |/| | |||||
| * | Add queue_coordinator / queue_syncer / gm. Shift to getting gen_server2 to ↵ | Simon MacMullen | 2013-12-10 | 1 | -1/+2 |
| | | | | | | | | store the name in most cases. | ||||
* | | simplify even further | Matthias Radestock | 2014-01-10 | 1 | -4/+4 |
| | | |||||
* | | simplify, remove redundant code | Matthias Radestock | 2014-01-10 | 1 | -7/+4 |
|/ | |||||
* | More consistent naming. | Simon MacMullen | 2013-09-11 | 1 | -1/+1 |
| | |||||
* | Take GM's view of the live members each time we have a death message; don't ↵ | Simon MacMullen | 2013-09-02 | 1 | -5/+7 |
| | | | | assume the master will always be able to write updates to Mnesia. | ||||
* | Refresh branch from stable | Emile Joubert | 2013-07-31 | 1 | -2/+27 |
|\ | |||||
| * | Merge bug25670 | Simon MacMullen | 2013-07-29 | 1 | -1/+3 |
| |\ | |||||
| | * | Don't roll over to a lower GM view version than we have ourselvesbug25671 | Emile Joubert | 2013-07-19 | 1 | -1/+3 |
| | | | |||||
| * | | Slaves detect stale master pids on startup | Emile Joubert | 2013-07-19 | 1 | -1/+24 |
| |/ | |||||
* | | s/VMware/GoPivotal/g | Simon MacMullen | 2013-07-01 | 1 | -2/+2 |
|/ | |||||
* | Merge in default | Simon MacMullen | 2013-01-30 | 1 | -3/+9 |
|\ | |||||
| * | Update copyright 2013bug25343 | Emile Joubert | 2013-01-23 | 1 | -1/+1 |
| | | |||||
| * | reduce wrongnessbug25340 | Matthias Radestock | 2012-12-04 | 1 | -2/+1 |
| | | |||||
| * | Send messages immediately if we are not busy. | Simon MacMullen | 2012-12-04 | 1 | -2/+9 |
| | |