condor_fetchlog [-pool centralmanagerhostname[:portnumber]] [-master -startd -schedd -collector -negotiator -kbdd] machine-name subsystem[.extension]
condor_fetchlog contacts HTCondor running on the machine specified by machine-name, and asks it to return a log file from that machine. Which log file is determined from the subsystem[.extension] argument. The log file is printed to standard output. This command eliminates the need to remotely log in to a machine in order to retrieve a daemon's log file.
For security purposes of authentication and authorization, this command requires ADMINISTRATOR level of access.
The subsystem[.extension] argument is utilized to construct the log file's name. Without an optional .extension, the value of the configuration variable named subsystem_LOG defines the log file's name. When specified, the .extension is appended to this value.
Typical strings for the argument subsystem are as given as possible values of the predefined configuration variable $(SUBSYSTEM). Note that access to any additional logs can be enabled by simply specifying the path to each log in the configuration file with a configuration parameter named <NAME>_LOG, choosing an arbitrary unique name for each case.
A value for the optional .extension argument is typically one of the three strings:
<X>is substituted with the slot number.
condor_fetchlog head.example.com NEGOTIATOR
To get the condor_startd daemon's log from a host named execute.example.com from within the current pool:
condor_fetchlog execute.example.com STARTD
This command requested the condor_startd daemon's log from the condor_master. If the condor_master has crashed or is unresponsive, ask another daemon running on that computer to return the log. For example, ask the condor_startd daemon to return the condor_master's log:
condor_fetchlog -startd execute.example.com MASTER