github/repmgr
1
0
Fork 0
mirror of https://github.com/EnterpriseDB/repmgr.git synced 2026-03-25 16:16:29 +00:00
Code Issues Packages Projects Releases 1 Wiki Activity

Improve database connection failure logging

Browse Source 
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?
This commit is contained in:
Ian Barwick
2019-03-20 11:27:12 +09:00
parent a38e229e61
commit 5cbaff8d0a
6 changed files with 47 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

29
repmgrd-physical.c
View File

@@ -537,8 +537,6 @@ check_primary_status(int degraded_monitoring_elapsed)
if (PQstatus(new_primary_conn) != CONNECTION_OK)
{
close_connection(&new_primary_conn);
if (primary_node_id == UNKNOWN_NODE_ID)
{
log_warning(_("unable to determine a new primary node"));
@@ -546,8 +544,11 @@ check_primary_status(int degraded_monitoring_elapsed)
else
{
log_warning(_("unable to connect to new primary node %i"), primary_node_id);
log_detail("\n%s", PQerrorMessage(new_primary_conn));
}
close_connection(&new_primary_conn);
/* "true" to indicate repmgrd should continue monitoring in degraded state */
return true;
}
@@ -1604,6 +1605,8 @@ monitor_streaming_witness(void)
}
else
{
log_warning(_("unable to connect to primary"));
log_detail("\n%s", PQerrorMessage(primary_conn));
/*
* Here we're unable to connect to a primary despite having scanned all
* known nodes, so we'll grab the record of the node we think is primary
@@ -2760,12 +2763,16 @@ promote_self(void)
/* connection should stay up, but check just in case */
if (PQstatus(local_conn) != CONNECTION_OK)
{
log_warning(_("local database connection not available"));
log_detail("\n%s", PQerrorMessage(local_conn));
local_conn = establish_db_connection(local_node_info.conninfo, true);
/* assume node failed */
if (PQstatus(local_conn) != CONNECTION_OK)
{
log_error(_("unable to reconnect to local node"));
log_detail("\n%s", PQerrorMessage(local_conn));
/* XXX handle this */
return FAILOVER_STATE_LOCAL_NODE_FAILURE;
}
@@ -2869,17 +2876,23 @@ notify_followers(NodeInfoList *standby_nodes, int follow_node_id)
for (cell = standby_nodes->head; cell; cell = cell->next)
{
log_verbose(LOG_DEBUG, "intending to notify node %i... ", cell->node_info->node_id);
log_verbose(LOG_DEBUG, "intending to notify node %i...", cell->node_info->node_id);
if (PQstatus(cell->node_info->conn) != CONNECTION_OK)
{
log_debug("reconnecting to node %i... ", cell->node_info->node_id);
log_info(_("reconnecting to node \"%s\" (node ID: %i)..."),
cell->node_info->node_name,
cell->node_info->node_id);
cell->node_info->conn = establish_db_connection(cell->node_info->conninfo, false);
}
if (PQstatus(cell->node_info->conn) != CONNECTION_OK)
{
log_debug("unable to reconnect to %i ... ", cell->node_info->node_id);
log_warning(_("unable to reconnect to \"%s\" (node ID: %i)"),
cell->node_info->node_name,
cell->node_info->node_id);
log_detail("\n%s", PQerrorMessage(cell->node_info->conn));
continue;
}
@@ -3788,6 +3801,7 @@ reset_node_voting_status(void)
if (PQstatus(local_conn) != CONNECTION_OK)
{
log_error(_("reset_node_voting_status(): local_conn not set"));
log_detail("\n%s", PQerrorMessage(local_conn));
return;
}
reset_voting_status(local_conn);
@@ -3799,7 +3813,10 @@ check_connection(t_node_info *node_info, PGconn **conn)
{
if (is_server_available(node_info->conninfo) == false)
{
log_warning(_("connection to node %i lost"), node_info->node_id);
log_warning(_("connection to node \"%s\" (ID: %i) lost"),
node_info->node_name,
node_info->node_id);
log_detail("\n%s", PQerrorMessage(*conn));
PQfinish(*conn);
*conn = NULL;
}