github/repmgr
1
0
Fork 0
mirror of https://github.com/EnterpriseDB/repmgr.git synced 2026-03-26 00:26:30 +00:00
Code Issues Packages Projects Releases 1 Wiki Activity

Document repmgrd's role on each type of node. Add

Browse Source 
logging (when run in verbose mode) showing what
decisions repmgrd is making after it determines
what type of node it is.
This commit is contained in:
Greg Smith
2011-02-24 18:58:03 -05:00
parent f928edf897
commit 459f14b594
2 changed files with 36 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
README.rst
View File

@@ -524,12 +524,28 @@ First, register the master by typing on "node1"::
repmgr -f /var/lib/pgsql/repmgr/repmgr.conf --verbose master register repmgr -f /var/lib/pgsql/repmgr/repmgr.conf --verbose master register
Start the "standby" server. Then start the "standby" server.
Register the standby by typing on "node2":: You could now register the standby by typing on "node2"::
repmgr -f /var/lib/pgsql/repmgr/repmgr.conf --verbose standby register repmgr -f /var/lib/pgsql/repmgr/repmgr.conf --verbose standby register
However, you can instead start repmgrd::
repmgrd -f /var/lib/pgsql/repmgr/repmgr.conf --verbose > /var/lib/pgsql/repmgr/repmgr.log 2>&1
Which will automatically register your standby system. And eventually
you need repmgrd running anyway, to save lag monitoring information.
repmgrd will log the deamon activity to the listed file. You can
watch what it is doing with::
tail -f /var/lib/pgsql/repmgr/repmgr.log
Hit control-C to exit this tail command when you are done.
Monitoring and testing
----------------------
At this point, you have a functioning primary on "node1" and a functioning At this point, you have a functioning primary on "node1" and a functioning
standby server running on "node2". You can confirm the master knows standby server running on "node2". You can confirm the master knows
about the standby, and that it is keeping it current, by looking at about the standby, and that it is keeping it current, by looking at
@@ -955,19 +971,16 @@ the program.
Usage Usage
----- -----
repmgrd reads the ``repmgr.conf`` file in current directory, or as indicated with -f repmgrd reads the ``repmgr.conf`` file in current directory, or as
parameter. It checks if the standby is in repl_nodes and adds it if not. indicated with -f parameter. If run on a standby, it checks if that
standby is in ``repl_nodes`` and adds it if not.
Before you can run the repmgr daemon (repmgrd) you need to register a master Before you can run repmgrd you need to register a master in a cluster
and at least a standby in a cluster using the ``MASTER REGISTER`` and using the ``MASTER REGISTER`` command. If run on a master,
``STANDBY REGISTER`` commands. repmgrd will exit, as it has nothing to do on them yet. It is only
targeted at running on standby servers currently. If converting
For example, following last example and assuming that ``repmgr.conf`` is in postgres a former master into a standby, you will need to start repmgrd
home directory you will run this on the master:: in order to make it fully operational in its new role.
repmgr -f /home/postgres/repmgr.conf master register
and the same in the standby.
The repmgr daemon creates 2 connections: one to the master and another to the The repmgr daemon creates 2 connections: one to the master and another to the
standby. standby.
@@ -975,9 +988,9 @@ standby.
Lag monitoring Lag monitoring
-------------- --------------
The repmgrd (repmgr daemon) helps monitor a set of master and standby repmgrd helps monitor a set of master and standby servers. You can
servers. You can see which node is the current master, as well as how see which node is the current master, as well as how far behind each
far behind each is from current. is from current.
To look at the current lag between primary and each node listed To look at the current lag between primary and each node listed
in ``repl_node``, consult the ``repl_status`` view:: in ``repl_node``, consult the ``repl_status`` view::
@@ -994,7 +1007,6 @@ This view shows the latest monitor info from every node.
* time_lag: in seconds. How many seconds behind the master is this node. * time_lag: in seconds. How many seconds behind the master is this node.
Error codes Error codes
----------- -----------

10
repmgrd.c
View File

@@ -157,14 +157,11 @@ main(int argc, char **argv)
snprintf(repmgr_schema, MAXLEN, "%s%s", DEFAULT_REPMGR_SCHEMA_PREFIX, local_options.cluster_name); snprintf(repmgr_schema, MAXLEN, "%s%s", DEFAULT_REPMGR_SCHEMA_PREFIX, local_options.cluster_name);
printf("Establishing database connection\n");
log_info(_("%s Connecting to database '%s'\n"), progname, local_options.conninfo); log_info(_("%s Connecting to database '%s'\n"), progname, local_options.conninfo);
myLocalConn = establishDBConnection(local_options.conninfo, true); myLocalConn = establishDBConnection(local_options.conninfo, true);
/* should be v9 or better */ /* should be v9 or better */
log_info(_("%s connected to database, checking its state\n"), progname); log_info(_("%s Connected to database, checking its state\n"), progname);
pg_version(myLocalConn, standby_version); pg_version(myLocalConn, standby_version);
if (strcmp(standby_version, "") == 0) if (strcmp(standby_version, "") == 0)
{ {
@@ -203,8 +200,13 @@ main(int argc, char **argv)
checkNodeConfiguration(local_options.conninfo); checkNodeConfiguration(local_options.conninfo);
if (myLocalMode == STANDBY_MODE) if (myLocalMode == STANDBY_MODE)
{ {
log_info(_("%s Starting continuous standby node monitoring'\n"), progname);
MonitorCheck(); MonitorCheck();
} }
else
{
log_info(_("%s This is a primary node, program not needed here; exiting'\n"), progname);
}
/* Prevent a double-free */ /* Prevent a double-free */
if (primaryConn == myLocalConn) if (primaryConn == myLocalConn)