github/repmgr
1
0
Fork 0
mirror of https://github.com/EnterpriseDB/repmgr.git synced 2026-03-27 08:56:29 +00:00
Code Issues Packages Projects Releases 1 Wiki Activity

repmgrd: fixes to failover handling

Browse Source 
get_new_primary() returns NULL if no notification for the new primary has
been received, but the code was expecting it to return UNKNOWN_NODE_ID,
which was causing repmgrd to prematurely drop out of the new primary
detection loop if no notification had been received by the time the loop
started.

Also store the electoral term as a single row, single column table,
to ensure that all repmgrds see the same turn. It is then bumped
by the winning node after it gets promoted.

Various logging improvements.
This commit is contained in:
Ian Barwick
2017-11-08 14:19:52 +09:00
parent 7232187f4d
commit 79d21b516b
5 changed files with 134 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

88
dbutils.c
View File

@@ -3578,9 +3578,52 @@ delete_monitoring_records(PGconn *primary_conn, int keep_history)
/*
* node voting functions
*
* These are intended to run under repmgrd and rely on shared memory
* These are intended to run under repmgrd and mainly rely on shared memory
*/
int
get_current_term(PGconn *conn)
{
PGresult *res = NULL;
int term = -1;
res = PQexec(conn, "SELECT term FROM repmgr.voting_term");
if (PQresultStatus(res) != PGRES_TUPLES_OK)
{
log_error(_("unable to query repmgr.voting_term:\n %s"),
PQerrorMessage(conn));
PQclear(res);
return -1;
}
term = atoi(PQgetvalue(res, 0, 0));
PQclear(res);
return term;
}
void
increment_current_term(PGconn *conn)
{
PGresult *res = NULL;
res = PQexec(conn, "UPDATE repmgr.voting_term SET term = term + 1");
if (PQresultStatus(res) != PGRES_COMMAND_OK)
{
log_error(_("unable to increment repmgr.voting_term:\n %s"),
PQerrorMessage(conn));
PQclear(res);
return;
}
PQclear(res);
return;
}
NodeVotingStatus
get_voting_status(PGconn *conn)
{
@@ -3700,19 +3743,26 @@ request_vote(PGconn *conn, t_node_info *this_node, t_node_info *other_node, int
}
int
set_voting_status_initiated(PGconn *conn)
void
set_voting_status_initiated(PGconn *conn, int electoral_term)
{
PQExpBufferData query;
PGresult *res = NULL;
int electoral_term = 0;
res = PQexec(conn, "SELECT repmgr.set_voting_status_initiated()");
initPQExpBuffer(&query);
appendPQExpBuffer(&query,
"SELECT repmgr.set_voting_status_initiated(%i)",
electoral_term);
res = PQexec(conn, query.data);
termPQExpBuffer(&query);
electoral_term = atoi(PQgetvalue(res, 0, 0));
PQclear(res);
return electoral_term;
return;
}
@@ -3748,7 +3798,6 @@ notify_follow_primary(PGconn *conn, int primary_node_id)
PQExpBufferData query;
PGresult *res = NULL;
initPQExpBuffer(&query);
appendPQExpBuffer(&query,
@@ -3756,10 +3805,17 @@ notify_follow_primary(PGconn *conn, int primary_node_id)
primary_node_id);
log_verbose(LOG_DEBUG, "notify_follow_primary():\n %s", query.data);
/* XXX handle failure */
res = PQexec(conn, query.data);
termPQExpBuffer(&query);
if (PQresultStatus(res) != PGRES_TUPLES_OK)
{
log_error(_("unable to execute repmgr.notify_follow_primary():\n %s"),
PQerrorMessage(conn));
PQclear(res);
return;
}
if (PQresultStatus(res) != PGRES_TUPLES_OK)
{
log_error(_("unable to execute repmgr.notify_follow_primary():\n %s"),
@@ -3786,16 +3842,24 @@ get_new_primary(PGconn *conn, int *primary_node_id)
res = PQexec(conn, query.data);
termPQExpBuffer(&query);
/* XXX handle error */
new_primary_node_id = atoi(PQgetvalue(res, 0, 0));
if (new_primary_node_id == UNKNOWN_NODE_ID)
if (PQresultStatus(res) != PGRES_TUPLES_OK)
{
log_error(_("unable to execute repmgr.reset_voting_status():\n %s"),
PQerrorMessage(conn));
PQclear(res);
return false;
}
if (PQgetisnull(res, 0, 0))
{
*primary_node_id = UNKNOWN_NODE_ID;
PQclear(res);
return false;
}
new_primary_node_id = atoi(PQgetvalue(res, 0, 0));
PQclear(res);
*primary_node_id = new_primary_node_id;