As the witness server does not, by definition, ever have an entry in pg_stat_replication,
we need to check its "attached" status by connecting to the witness server itself
and querying the reported upstream node ID (which should be set by the witness
server repmgrd). If this matches the current primary node ID, we count it as attached.
If the sibling node has already been promoted (for whatever
reason, e.g. "repmgr standby promote" was executed manually)
and has exited recovery, the upstream node ID will normally
be reported as "-1", which is correct, but looks confusing in
the logs.
We now only report the upstream node ID if the sibling node
is still in recovery, *or* if it has exited recovery but is
still reporting an extant node ID.
This functionality enables repmgrd (when running on the primary) to
monitor connected child nodes. It will log connections and disconnections
and generate events.
Additionally, repmgrd can execute a custom script if the number of connected
child nodes falls below a configurable threshold. This script can be used
e.g. to "fence" the primary following a failover situation where a new primary
has been promoted and all standbys are now child nodes of that primary.
Unless the PQExpBuffer is required for the duration of the function,
ensure it's always a variable local to the relevant code block. This
mitigates the risk of accidentally accessing a generically named
PQExpBuffer which hasn't been initialised or was previously terminated.
Previously, repmgrd assumed that during a failover, there would not
already be another primary node. However it's possible a node was
promoted manually. While this is not a desirable situation, it's
conceivable this could happen in the wild, so we should check for
it and react accordingly.
Also sanity-check that the follow target can actually be followed.
Addresses issue raised in GitHub #420.
Log the output of PQerrorStatus() in a couple of places where it was missing.
Additionally, always log the output of PQerrorStatus() starting with a blank
line, otherwise the first line looks like it was emitted by repmgr, and
it's harder to scan the error message.
Before:
[2019-03-20 11:24:15] [DETAIL] could not connect to server: Connection refused
Is the server running on host "localhost" (::1) and accepting
TCP/IP connections on port 5501?
could not connect to server: Connection refused
Is the server running on host "localhost" (127.0.0.1) and accepting
TCP/IP connections on port 5501?
After:
[2019-03-20 11:27:21] [DETAIL]
could not connect to server: Connection refused
Is the server running on host "localhost" (::1) and accepting
TCP/IP connections on port 5501?
could not connect to server: Connection refused
Is the server running on host "localhost" (127.0.0.1) and accepting
TCP/IP connections on port 5501?
If "failover_validation_command" is set, and the command returns an error,
rerun the election.
There is a pause between reruns to avoid "churn"; the length of this pause
is controlled by the configuration parameter "election_rerun_interval".
This controls the maximum length of time in seconds that repmgrd will
wait for other standbys to disconnect their WAL receivers in a failover
situation.
This setting is only used when "standby_disconnect_on_failover" is set to "true".
If the WAL receiver has been temporarily disabled, we don't want to
wait for it to start up as it may not be able to at that point; we do
however need to reset "wal_retrieve_retry_interval".
"standby_disconnect_on_failover" requires availability of "wal_retrieve_retry_interval",
which is available from PostgreSQL 9.5.
9.4 will fall out of community support this year, so it doesn't seem
productive at this point to do anything more than put the onus on the user
to read the documentation and heed any warning messages in the logs.
This is intended to ensure that all nodes have a constant LSN while
making the failover decision.
This feature is experimental and needs to be explicitly enabled with the
configuration file option "standby_disconnect_on_failover".
Note enabling this option will result in a delay in the failover decision
until the WAL receiver is disconnected on all nodes.
This enable selection of the method repmgrd uses to check whether the upstream
node is available. Possible values are:
- "ping" (default): uses PQping() to check server availability
- "connection": executes a query on the connection to check server
availability (similar to repmgr3.x).
In a situation where only some standbys are cut off from the primary,
a failover would result in a split brain/split cluster situation,
as it's likely one of the cut-off standbys will promote itself, and
other cut-off standbys (but not all standbys) will follow it.
To prevent this happening, interrogate the other sibiling nodes to
check whether they've seen the primary within a reasonably short interval;
if this is the case, do not take any failover action.
This feature is experimental.
If, for whatever reason, repmgrd is not running on a node, but that
node qualifies as promotion candidate, failover will not take place
as that node will never promote itself.
We therefore discount nodes where repmgrd is running as promotion
candidates, which will ensure one node is always promoted.
There is a slight risk here that the node(s) where repmgrd is not running
are further ahead, leading to a timeline fork. It might be possible
to mitigate that by having the "election" leader perform the promote
(or follow) operation.
Ian Barwick