github/repmgr
1
0
Fork 0
mirror of https://github.com/EnterpriseDB/repmgr.git synced 2026-03-23 07:06:30 +00:00
Code Issues Packages Projects Releases 1 Wiki Activity

repmgrd: fix startup on witness node when local data is stale

Browse Source 
Previously, when running on a witness server, repmgrd didn't consider
the local cache of the "repmgr.nodes" table might be outdated, e.g.
as repmgrd wasn't running on the witness server during a failover,
so could potentially end up monitoring a former primary now running
as a standby.

When running on a witness server, at startup repmgrd will now scan
all nodes to determine the current primary, and refresh its local
cache from there. This will also ensure it can start up even if the
node currently registered as primary in the local cache is not available.

Implements GitHub #488 and #489.
This commit is contained in:
Ian Barwick
2018-08-09 16:36:18 +09:00
parent 987823861f
commit 221fb63e92
1 changed files with 9 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

36
repmgrd-physical.c
View File

@@ -1173,36 +1173,18 @@ monitor_streaming_witness(void)
PQExpBufferData event_details;
RecordStatus record_status;
int primary_node_id = UNKNOWN_NODE_ID;
reset_node_voting_status();
log_debug("monitor_streaming_witness()");
if (get_primary_node_record(local_conn, &upstream_node_info) == false)
{
PQExpBufferData event_details;
initPQExpBuffer(&event_details);
appendPQExpBuffer(&event_details,
_("unable to retrieve record for primary node"));
log_error("%s", event_details.data);
log_hint(_("execute \"repmgr witness register --force\" to update the witness node "));
close_connection(&local_conn);
create_event_notification(NULL,
&config_file_options,
config_file_options.node_id,
"repmgrd_shutdown",
false,
event_details.data);
termPQExpBuffer(&event_details);
terminate(ERR_BAD_CONFIG);
}
primary_conn = establish_db_connection(upstream_node_info.conninfo, false);
/*
* At this point we can't trust the local copy of "repmgr.nodes", as
* it may not have been updated. We'll scan the cluster for the current
* primary and refresh the copy from that before proceeding further.
*/
primary_conn = get_primary_connection_quiet(local_conn, &primary_node_id, NULL);
/*
* Primary node must be running at repmgrd startup.
@@ -1227,7 +1209,7 @@ monitor_streaming_witness(void)
* refresh upstream node record from primary, so it's as up-to-date
* as possible
*/
record_status = get_node_record(primary_conn, upstream_node_info.node_id, &upstream_node_info);
record_status = get_node_record(primary_conn, primary_node_id, &upstream_node_info);
/*
* This is unlikely to happen; if it does emit a warning for diagnostic