You can control the manager function with the dscontrol manager command.
>>-dscontrol--manager--+-interval--seconds-------------------------------------------------+->< +-loglevel--level---------------------------------------------------+ +-logsize--+-unlimited-+--------------------------------------------+ | '-bytes-----' | +-metric set--+-loglevel--level--------+----------------------------+ | '-logsize--+-unlimited-+-' | | '-bytes-----' | +-quiesce--+-server--------------+--+-----------------------------+-+ | '-cluster@port@server-' '-daily--start_hour--end_hour-' | +-reach set--+-interval--seconds------+-----------------------------+ | +-loglevel--level--------+ | | '-logsize--+-unlimited-+-' | | '-bytes-----' | +-refresh--refresh cycle--------------------------------------------+ +-report--+---------------------+-----------------------------------+ | '-cluster+cluster2...-' | +-restart--Message--------------------------------------------------+ +-sensitivity--weight-----------------------------------------------+ +-smoothing--index--------------------------------------------------+ +-start----log file--metric_port------------------------------------+ +-status------------------------------------------------------------+ +-stop--------------------------------------------------------------+ +-unquiesce--+-server--------------+--------------------------------+ | '-cluster@port@server-' | '-version-----------------------------------------------------------'
A positive number representing in seconds how often the manager will update weights to the executor. The default is 2.
The number of the level (0 to 5). The higher the number, the more information that is written to the manager log. The default is 1. The following are the possible values: 0 is None, 1 is Minimal, 2 is Basic, 3 is Moderate, 4 is Advanced, 5 is Verbose.
The maximum size in bytes for the manager log file. You can specify either a positive number greater than zero, or the word unlimited. The log file may not reach the exact maximum size before overwriting because the log entries themselves vary in size. The default value is 1 MB.
The IP address of the server as either a symbolic name or in dotted decimal format.
When you want to bring down a particular service on a server, like HTTP, you can now quiesce individual daemons. If you have added server1 to port 80 and port 22 of a cluster, for example, the usual quiesce command will stop forwarding traffic to server1 on port 80 and port 22. This new option allows you to stop traffic only on port 80 if you need to stop that particular service.
A positive number representing the number of intervals. The default is 2.
A message that you want written to the manager log file.
A number from 1 to 100 to be used as the weight percentage. The default of 5 creates a minimum sensitivity of 5%.
A positive floating point number. The default is 1.5.
File name to which the manager data is logged. Each record in the log is time stamped. The default file is installed in the logs directory. See Examples. To change the directory where the log files are kept, see Logging with Load Balancer.
Port that Metric Server will use to report system loads. If you specify a metric port, you must specify a log file name. The default metric port is 10004.
The IP address of the server as either a symbolic name or in dotted decimal format.
This command will unquiesce that particular port for service, but all other ports will remained quiesced.
dscontrol manager interval 5
dscontrol manager loglevel 0
dscontrol manager logsize 1000000
dscontrol manager quiesce 130.40.52.153
dscontrol manager quiesce cluster1@9.40.25.67@80
dscontrol manager refresh 3
dscontrol manager report
dscontrol manager restart Restarting the manager to update code
dscontrol manager sensitivity 10
dscontrol manager smoothing 2.0
dscontrol manager start ndmgr.log
dscontrol manager status
dscontrol manager stop
dscontrol manager quiesce 130.40.52.153 daily 2 16
dscontrol manager unquiesce 130.40.52.153
dscontrol manager version