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: improve reconnection handling

Browse Source 
Previously, if the server being monitored was not available, repmgrd
would always close the existing connection handle and open a new one.

However, in some cases, e.g. a brief network outage, the existing
connection handle is still good and does not need to be reopened.

This could be particularly problematic if monitoring_history is on,
as this risks leaving orphan sessions on the primary which (given
a sufficiently unstable network) could lead to all available backends
being occupied.

Instead, during an outage we now use a new connection to verify
the server is accessible; if the old connection is still available
(e.g. following a short network interruption) we continue using that;
if  not (e.g. the server was restarted), we use the new one.
This commit is contained in:
Ian Barwick
2018-08-30 10:24:06 +09:00
parent 216326f316
commit 0468e47ef3
8 changed files with 59 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
repmgrd-physical.c
View File

@@ -288,8 +288,6 @@ monitor_streaming_primary(void)
local_node_info.node_status = NODE_STATUS_UNKNOWN;
close_connection(&local_conn);
/*
* as we're monitoring the primary, no point in trying to
* write the event to the database
@@ -305,7 +303,7 @@ monitor_streaming_primary(void)
termPQExpBuffer(&event_details);
local_conn = try_reconnect(&local_node_info);
try_reconnect(&local_conn, &local_node_info);
if (local_node_info.node_status == NODE_STATUS_UP)
{
@@ -744,8 +742,6 @@ monitor_streaming_standby(void)
log_warning("%s", event_details.data);
termPQExpBuffer(&event_details);
close_connection(&upstream_conn);
/*
* if local node is unreachable, make a last-minute attempt to reconnect
* before continuing with the failover process
@@ -756,7 +752,7 @@ monitor_streaming_standby(void)
check_connection(&local_node_info, &local_conn);
}
upstream_conn = try_reconnect(&upstream_node_info);
try_reconnect(&upstream_conn, &upstream_node_info);
/* Node has recovered - log and continue */
if (upstream_node_info.node_status == NODE_STATUS_UP)
@@ -1087,7 +1083,7 @@ loop:
* if monitoring not in use, we'll need to ensure the local connection
* handle isn't stale
*/
connection_ping(local_conn);
(void) connection_ping(local_conn);
}
/*
@@ -1304,8 +1300,7 @@ monitor_streaming_witness(void)
true,
event_details.data);
close_connection(&primary_conn);
primary_conn = try_reconnect(&upstream_node_info);
try_reconnect(&primary_conn, &upstream_node_info);
/* Node has recovered - log and continue */
if (upstream_node_info.node_status == NODE_STATUS_UP)