github/repmgr
1
0
Fork 0
mirror of https://github.com/EnterpriseDB/repmgr.git synced 2026-03-26 16:46:28 +00:00
Code Issues Packages Projects Releases 1 Wiki Activity

repmgrd: check for a change to the upstream node

Browse Source 
If the upstream node has changed, for example after "repmgr standby follow"
was manually executed, restart monitoring to ensure repmgrd is monitoring the
correct node.
This commit is contained in:
Ian Barwick
2019-01-22 13:31:00 +09:00
parent b6fe91ebcd
commit 1980deb480
3 changed files with 99 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

57
repmgrd-physical.c
View File

@@ -630,7 +630,9 @@ check_primary_status(int degraded_monitoring_elapsed)
}
/*
* repmgrd running on a standby server
*/
void
monitor_streaming_standby(void)
{
@@ -641,6 +643,8 @@ monitor_streaming_standby(void)
MonitoringState local_monitoring_state = MS_NORMAL;
instr_time local_degraded_monitoring_start;
int last_known_upstream_node_id = UNKNOWN_NODE_ID;
log_debug("monitor_streaming_standby()");
reset_node_voting_status();
@@ -729,6 +733,8 @@ monitor_streaming_standby(void)
return;
}
last_known_upstream_node_id = local_node_info.upstream_node_id;
/*
* refresh upstream node record from upstream node, so it's as up-to-date
* as possible
@@ -875,7 +881,6 @@ monitor_streaming_standby(void)
if (PQstatus(local_conn) != CONNECTION_OK)
{
check_connection(&local_node_info, &local_conn);
}
@@ -1144,7 +1149,7 @@ monitor_streaming_standby(void)
*/
/* refresh own internal node record */
record_status = get_node_record(local_conn, local_node_info.node_id, &local_node_info);
record_status = refresh_node_record(local_conn, local_node_info.node_id, &local_node_info);
if (local_node_info.type == PRIMARY)
{
@@ -1354,20 +1359,48 @@ loop:
}
else
{
int stored_local_node_id = repmgrd_get_local_node_id(local_conn);
int stored_local_node_id = UNKNOWN_NODE_ID;
if (local_monitoring_state == MS_DEGRADED)
{
log_info(_("connection to local node recovered after %i seconds"),
calculate_elapsed(local_degraded_monitoring_start));
local_monitoring_state = MS_NORMAL;
/*
* Check if anything has changed since the local node came back on line;
* we may need to restart monitoring.
*/
refresh_node_record(local_conn, local_node_info.node_id, &local_node_info);
if (last_known_upstream_node_id != local_node_info.upstream_node_id)
{
log_notice(_("local node %i upstream appears to have changed, restarting monitoring"),
local_node_info.node_id);
log_detail(_("currently monitoring upstream %i; new upstream is %i"),
last_known_upstream_node_id,
local_node_info.upstream_node_id);
close_connection(&upstream_conn);
return;
}
/*
*
*/
if (local_node_info.type != STANDBY)
{
log_notice(_("local node %i is no longer a standby, restarting monitoring"),
local_node_info.node_id);
close_connection(&upstream_conn);
return;
}
}
/*
* If the local node was restarted, we'll need to reinitialise values
* stored in shared memory.
*/
stored_local_node_id = repmgrd_get_local_node_id(local_conn);
if (stored_local_node_id == UNKNOWN_NODE_ID)
{
@@ -1428,9 +1461,23 @@ loop:
handle_sighup(&local_conn, STANDBY);
}
refresh_node_record(local_conn, local_node_info.node_id, &local_node_info);
if (local_monitoring_state == MS_NORMAL && last_known_upstream_node_id != local_node_info.upstream_node_id)
{
log_notice(_("local node %i's upstream appears to have changed, restarting monitoring"),
local_node_info.node_id);
log_detail(_("currently monitoring upstream %i; new upstream is %i"),
last_known_upstream_node_id,
local_node_info.upstream_node_id);
close_connection(&upstream_conn);
return;
}
log_verbose(LOG_DEBUG, "sleeping %i seconds (parameter \"monitor_interval_secs\")",
config_file_options.monitor_interval_secs);
sleep(config_file_options.monitor_interval_secs);
}
}