When EMQX boots up, it tries to get latest config from peer (core type)
nodes, if none of the nodes are replying, the node will decide
to boot with local config (and replay the committed changes) if
the commit table is loaded from disk locally (an indication of the
data being latest), otherwise it will sleep for 1-2 seconds and
retry.
This lead to a race condition, e.g. in a two nodes cluster:
1. node1 boots up
2. node2 boots up and copy mnesia table from node1
3. node1 restart before node2 can sync cluster.hocon from it
4. node1 boots up and copy mnesia table from node2
Now that both node1 and node2 has the mnesia `load_node` pointing
to each other (i.e. not a local disk load).
Prior to this fix, the nodes would wait for each other in a dead loop.
This commit fixes the issue by allowing node to boot
with local config if it does not have a lagging.
* feat(emqx_conf): move conf manager for emqx_machine to emqx_conf
* chore(emqx_conf): change emqx:get_config/2 to emqx_conf:get/2
* fix: common test failed
* fix: badmatch by typo wrong key
* fix(emqx_conf): get the wrong core nodes
* fix(emqx_conf): get core node's tnx_id not latest tnx_id
* fix: add ro_transation when copy conf file
* fix: delete debug info
* fix: change ekka_rlog to mria_rlog
* fix: remove cluster_rpc from emqx_machine.
* fix: don't call ekka:start/0 explicitly
* fix: ekka should be start in emqx_machine
Ivan Dyachkov