In case if it was interrupted by SIGCHLD but scheduled awake time is not
reached it will continue sleeping. For all other signals behaviuor is
not changed.
In order to have backward compatibility connection url of restapi is
passed as application_name parameter in connection url of database.
Value in etcd will look like:
postgres://user:passwd@host:5432/postgres?application_name=http://host:8080/governor
Sure, this is a hack, but I guess it worth it.
Server always returns json which contains some status of postgres:
* instance type: master/slave
* xlog location for master
* xlog received_location, replayed_location
This server could be used by haproxy
GET / returns 200 if there is postgres master behind governor.
GET /slave returns 200 is there is postgres slave behind governor
In all other cases it returns 503
Currently server always listening on 0.0.0.0:8080, but it should be
configurable. In the next versions this rest api could also report some
status of etcd and could be used by is_healthiest_node method
Refactored query and is_healthiest_node methods
query method wont retry more then 3 times if OperationalError raised
is_healthiest_node takes into account that one of the nodes could be
still a master
For example you can listen on 0.0.0.0 but provide service on some
didicated ip. This should cover all possible use-cases when you are
running behind firewall, load balancer or inside for example docker.
Basically this is rollback to the original decision tree with the small exception:
1 - If leader is defined and it's not me - then slave would be started immidiately
with the correct recovery conf.
2 - If leader is defined in and it's my host, then it will start instance in
readonly (but without primary_conninfo in recovery.conf)
3 - And the third case - if the leader is not defined - it also will start
instance in readonly, without primary_conninfo.
After performing 2 or 3 it will perform usual decision tree.
Also governor is able to pick up already running master and slave
instances without restarting them.
In case is you have a lock and postgres is not running behaviour remains
the same: it will start master in read only mode and then promote if it
still has the lock.
Alexander Kukushkin