summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* merge defaultFrancesco Mazzoli2012-08-139-816/+1246
|\
| * merge defaultFrancesco Mazzoli2012-08-0121-224/+670
| |\
| * | add a first draft for the rabbitmqctl documentationFrancesco Mazzoli2012-07-252-57/+138
| | | | | | | | | | | | Also, accept disk as well as disc in `change_cluster_node_type'.
| * | merge defaultFrancesco Mazzoli2012-07-168-121/+119
| |\ \
| * | | clearer `rabbit_mnesia:discover_cluster/1'Francesco Mazzoli2012-07-091-4/+6
| | | |
| * | | `cluster_status/{0,1}' now returns the status directly, fails on errorFrancesco Mazzoli2012-07-091-11/+7
| | | |
| * | | cosmeticFrancesco Mazzoli2012-07-091-16/+17
| | | |
| * | | do not update the cluster status file by getting the mnesia statusFrancesco Mazzoli2012-07-091-8/+4
| | | | | | | | | | | | | | | | | | | | | | | | Instead, just update the existing cluster status file. This is because the mnesia propagation time can be slower than the other messages (see other comments in the bug)
| * | | add the magic `cluster' word to `remove_node' and `change_node_type'Francesco Mazzoli2012-07-092-15/+15
| | | |
| * | | `rabbit_misc:rabbit_version/0' => `rabbit_misc:version/0'Francesco Mazzoli2012-07-092-5/+5
| | | |
| * | | get rid of `is_disc_node'Francesco Mazzoli2012-07-091-9/+5
| | | |
| * | | remove useless `all_clustered_nodes/0' in `reset/1'Francesco Mazzoli2012-07-091-2/+2
| | | |
| * | | split the offline case of `rabbit_mnesia/2' in a separate functionFrancesco Mazzoli2012-07-091-49/+51
| | | |
| * | | `change_extra_db_nodes/2' throws instead of returning errorFrancesco Mazzoli2012-07-091-43/+39
| | | |
| * | | remove unnecessary `ensure_mnesia_not_running/0'Francesco Mazzoli2012-07-091-2/+1
| | | |
| * | | ?SERVER instead of explicit nameFrancesco Mazzoli2012-07-091-3/+3
| | | |
| * | | check that the node selected for removal is in the clusterFrancesco Mazzoli2012-07-091-1/+1
| | | |
| * | | fix the flags declaration in `rabbit_control_main'Francesco Mazzoli2012-07-061-1/+1
| | | |
| * | | correct invocations of `on_node_{up,down}' in `rabbit_node_monitor'.Francesco Mazzoli2012-07-061-2/+3
| | | |
| * | | Merge headsSimon MacMullen2012-07-061-1/+1
| |\ \ \
| | * | | While this didn't actually crash the VM for me, we generally want to HiPE as ↵Simon MacMullen2012-07-061-1/+1
| | | | | | | | | | | | | | | | | | | | early as we possibly can.
| * | | | merge defaultFrancesco Mazzoli2012-07-0526-113/+589
| |\ \ \ \
| * | | | | fix `remove_node/1' - do not remove nodes from offline nodes, unless we're ↵Francesco Mazzoli2012-07-052-37/+81
| | |/ / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | "safe" We wan to remove nodes from offline nodes only when all the cluster nodes are offline, and when the node we are removing from was either the last to go down or the second after the node we are removing. However we cannot know that for sure. For this reason rabbit will refuse to remove from offline nodes unless the flag `--offline' is passed. There is more information in comments around `rabbit_mnesia:remove_node/2.
| * | | | fix `recluster/1' - delete the schema and write status before `init_db'Francesco Mazzoli2012-07-051-7/+10
| | | | |
| * | | | don't show the running nodes when downFrancesco Mazzoli2012-07-041-2/+5
| | | | |
| * | | | check if the node is the only disc one in `change_node_type/1'Francesco Mazzoli2012-07-041-10/+16
| | | | |
| * | | | check if only disc node after `init_db' in `reset/1'Francesco Mazzoli2012-07-041-10/+9
| | | | |
| * | | | do not check for consistency when resetingFrancesco Mazzoli2012-07-041-10/+20
| | | | |
| * | | | reworked functions that get cluster nodes to get only what they needFrancesco Mazzoli2012-07-041-26/+51
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This is mainly do avoid a deadlock caused by running `rabbit:stop/1', which indirectly calls `applications:which_applications/1', from inside `application:stop'.
| * | | | all_clustered_ram_nodes() => clustered_ram_nodes()Francesco Mazzoli2012-07-041-2/+2
| | | | |
| * | | | wipe the mnesia data before starting if the cluster nodes are inconsistentFrancesco Mazzoli2012-07-033-38/+74
| | | | |
| * | | | better `rabbit_mnesia:check_cluster_consistency/0'Francesco Mazzoli2012-07-031-28/+21
| | | | |
| * | | | merge defaultFrancesco Mazzoli2012-06-285-7/+27
| |\ \ \ \
| * | | | | `cluster_status_from_mnesia/0' fails if tables are not presentFrancesco Mazzoli2012-06-281-19/+30
| | | | | |
| * | | | | don't use `system_info(running_db_nodes)', check for running rabbits insteadFrancesco Mazzoli2012-06-281-3/+17
| | | | | |
| * | | | | instead of manually removing the current node, filter it in `cluster_status'Francesco Mazzoli2012-06-283-17/+9
| | | | | |
| * | | | | moved the status file handling to `rabbit_node_monitor', status from mnesia ↵Francesco Mazzoli2012-06-273-275/+320
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | when online I'm much happier with this configuration, with the exception of one problem: `rabbit_mnesia:cluster_status' will happily return the data from mnesia whenever mnesia is online, but that gives wrong results when the node is a RAM node and it just got online. I have some ideas to improve this.
| * | | | | change semantics of the cluster status operationsFrancesco Mazzoli2012-06-271-55/+54
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Now they use the file only when offline. Also I change the naming, from a `cluster_nodes_status' suffix to `cluster_status_file'.
| * | | | | change `rabbit_mnesia:cluster_status_if_running' slightlyFrancesco Mazzoli2012-06-271-8/+12
| | | | | |
| * | | | | check if only running disc node only in the required placesFrancesco Mazzoli2012-06-271-4/+6
| | | | | |
| * | | | | check that the node is in the cluster in `rabbit_mnesia:remove_node/1'Francesco Mazzoli2012-06-261-0/+5
| | | | | |
| * | | | | remove the node from the running node when neededFrancesco Mazzoli2012-06-261-18/+23
| | | | | |
| * | | | | notify cluster of new nodes as soon as they join, and not when they startFrancesco Mazzoli2012-06-262-14/+38
| | | | | |
| * | | | | merge defaultFrancesco Mazzoli2012-06-228-720/+867
| |\ \ \ \ \
| | * \ \ \ \ merge defaultFrancesco Mazzoli2012-06-208-30/+23
| | |\ \ \ \ \
| | * | | | | | add a "leave cluster" hook in `rabbit_node_monitor', handle cluster nodes ↵Francesco Mazzoli2012-06-142-27/+66
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | joins/departures better
| | * | | | | | merge defaultFrancesco Mazzoli2012-06-147-707/+815
| | |\ \ \ \ \ \
| | | * \ \ \ \ \ merge defaultFrancesco Mazzoli2012-06-147-707/+815
| | | |\ \ \ \ \ \
| | | | * \ \ \ \ \ merge defaultFrancesco Mazzoli2012-06-128-106/+94
| | | | |\ \ \ \ \ \
| | | | * | | | | | | revert last commitFrancesco Mazzoli2012-06-124-77/+40
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | I had completely misunderstood what fhc does. I'll put the atomic changes to the cluster status file in the `rabbit_node_monitor'.
View Lorry Controller Status