# pxc-scheduler-handler configuration

*ProxySQL* with the `pxc-scheduler-handler` stores the parameters in a
configuration file that uses the toml format. This file defines
the server credentials and other settings.

Use the `--config-file` option to run the pxc-scheduler-handler script.

!!! warning

    We do not recommend storing the scheduler configuration file in the Home directory.

## Example of a configuration file

A configuration file is used to control the `pxc-scheduler-handler` operations. The file can be changed as needed. The administrator controls what protected resources the tool can access. 

```{.text .no-copy}
# For the detailed manual, see
# https://github.com/percona/pxc_scheduler_handler#how-to-configure-pxc-scheduler-handler


[pxccluster]
activeFailover = 1
failBack = false
checkTimeOut = 2000
pingTimeout = 1000
mainSegment = 0
sslClient = "client-cert.pem"
sslKey = "client-key.pem"
sslCa = "ca.pem"
sslCertificatePath = "/path/to/ssl_cert"
hgW = 100
hgR = 101
configHgRange =8000
maintenanceHgRange =9000

# --------------------------------
# Set to true if there is a single writer node.  If this is set,
# then maxNumWriters is assumed to be 1.
#
# Allowable values: true,false
# Default: false
#
singlePrimary = true

# --------------------------------
# Set to the number of writer nodes desired.
#
# The value of this is assumed to be 1 if singlePrimary is true.
#
# If this is set to a value from 1 to 100, then the query rules
# are setup for a distinct writer hostgroup (writes are sent to the
# writer hostgroup and read are sent to the reader hostgroup).
#
# If this is set to a value > 100, then all queries (writes and reads)
# are sent to the writer hostgroup.  This is assumed to be a
# load-balancing scenario, where all nodes are equivalent and accept
# both reads and writes.
#
# Default: (none)
#
maxNumWriters = 1
writerIsAlsoReader = 1
retryUp = 0
retryDown = 2
clusterId = 10

# Controls the primary settings during failover.
# More details at https://github.com/percona/pxc_scheduler_handler#persist-primary-values
#
# Allowed values:
#
#       0 Disable
#       1 Persist only write settings
#       2 Persist both read and write settings
persistPrimarySettings=0



# == proxysql ===================================================
# The proxysql section is for ProxySQL-specific information.
#
# These settings will be read and used whenever the scheduler is run.
#
[proxysql]
port = 6032
host = "127.0.0.1"
user = "<valid user to connect from real ip as for proxysql_server table>"
password = "<password>"
clustered = false
lockfilepath ="/var/run/pxc_scheduler_handler"
respectManualOfflineSoft=false
pingTimeout = 1000



#== global ======================================================
# The global section are for variables that are not ProxySQL or
# cluster specific.
#
# These settings will be read and used whenever the scheduler is run.
#
[global]
debug = true

# stdout: output is redirected to proxysql logs
# file: output is written to the file pointed by logFile. When setting 
to ``file`, ensure that the user `proxysql:proxysql` has the 
appropriate permissions to write to this location.
logTarget = "stdout" #stdout | file

# Defines the log level to be used.
# Allowed options are [error,warning,info,debug]
logLevel = "info"
logFile = "/var/log/pxc_scheduler_handler/pscheduler.log"

# Should be set to false if we are pxc_scheduler_handler through percona-scheduler-admin.
daemonize = false
daemonInterval = 2000

# boolean variable which enables reporting of statistics.
performance = true

# Not used currently
OS = "na"

# Time in seconds after which the file lock is considered expired [local instance lock]
lockFileTimeout = 60 #seconds

# Time in seconds after which the cluster lock is considered expired
lockClusterTimeout = 600 #seconds



#== setup =======================================================
# These variables are used only upon Setup
# Changing these variables after setup will not affect operation
#
[setup]

# --------------------------------
# The clusterAppUser is the ProxySQL user account that should be
# used by clients to access the cluster.
#
# Uncomment the following options (clusterAppUser and clusterAppUserPassword)
# to enable the setting of the clusterAppUser for this cluster.
#
#clusterAppUser="proxysql_user"
#clusterAppUserPassword="passw0rd"

# --------------------------------
# The monitorUser is used by ProxySQL to access the servers and
# check the connections.
#
monitorUser="monitor"
monitorUserPassword="monitor"

# --------------------------------
# The clusterXXX information is used to setup the cluster for
# use by ProxySQL.
#
clusterHost="<IP_ADDRESS>"
clusterPort=3306
clusterUser="admin"
clusterUserPassword="admin"

# --------------------------------
# ProxySQL will use SSL to connect to the backend servers
#
useSSL=0

# --------------------------------
# Max number of connections from ProxySQL to the backend servers.
#
maxConnections=1000


# --------------------------------
# Defines how frequently (in milliseconds) the scheduler must be run
#
nodeCheckInterval=2000
```