Job Control Environment Variables (JOB CONTROL IS DEPRECATED)
There are a number of environment variables that can be used to specify the location of resources, manage the job database, and control various aspects of job execution. Some of these environment variables provide alternate means of specifying the resource; others are the sole means of specifying a resource or of overriding a default value.
|
Variable |
Purpose |
||||||
|
|
Specify the license file or a list of license files to use if the license file is not installed in |
||||||
|
|
Specify the installation directory. |
||||||
|
|
If set to 1, force the contents of the job directory to be archived and copied back to the submission host only for those jobs that have died (whose exit status is “died”). It is also possible to force such copying for all jobs (see
|
||||||
|
|
Time threshold after which an active but unreachable job is considered dead. |
||||||
|
|
Specify the hosts file. See Location of the Hosts File for a Job. |
||||||
|
|
Control debugging output. Allowed values are:
|
||||||
|
|
Specify how to incorporate job results into the project (same as entering Available values are:
Default: |
||||||
|
|
Specify the full path name for the job database. |
||||||
|
|
Specify how long the job record of a completed job is kept in the job database before being automatically deleted. For example, one day can be expressed as See Purging the Job Database (JOB CONTROL IS DEPRECATED) and Performing Actions on Jobs with jnanny (JOB CONTROL IS DEPRECATED). |
||||||
|
|
Specify the maximum number of job records to be kept in the job database. If this number is exceeded, job records of completed jobs are purged, starting with the oldest, at the next cleanup time. Default: 2000 records. |
||||||
|
|
Specify the port or range of ports that |
||||||
|
|
Regulate the use of |
||||||
|
|
Specify the port or range of ports that |
||||||
|
|
Regulate the use of |
||||||
|
|
Set the size threshold for the |
||||||
|
|
Specify the port or range of ports that |
||||||
|
|
Turns on debugging related to license usage. Any nonzero integer value turns on debugging. |
||||||
|
|
Specify how long to keep trying to obtain a license. The time interval between license checkout attempts can either be doubling, or constant. To use doubling time intervals, provide an integer to specify the time in seconds, or an integer value with a time unit, such as 7200s, 120m, 2h. An attempt to obtain a license is made at intervals that double after each try, starting at 10s. The final attempt is made at the specified maximum time. To use a constant time interval between license checkout attempts, provide a colon-separated pair of a time and an integer. For example,
Default: 1 minute. |
||||||
|
|
Clean up (remove) Default: 1 month. |
||||||
|
|
Specify the number of times a failed subjob from a distributed job is rerun before exiting with a failure. Only supported for distributed Glide and workflow jobs. Default: 3. |
||||||
|
|
Set the interval at which monitored files are copied back to the launch directory. |
||||||
|
|
Lower the priority of Schrödinger jobs. If it is unset or set to a null string, the priority is not lowered. The value Equivalent to using the |
||||||
|
|
If set, avoid using NFS mounts for file transfers between the job directory and the submission (output) directory, if possible. |
||||||
|
|
Specify the path to a file containing a list of hosts and the number of CPUs to use on each host. Each line in the file must have the format: You can list a host multiple times rather than provide the number of CPUs; this may affect the execution of parallel jobs. |
||||||
|
|
This environment variable is used to pass an explicit list of hosts to the driver for a parallel or distributed job. It is not used by Job Control directly; as such, the meaning of the file depends on the product that uses it. For a parallel job, such as parallel Jaguar or Desmond, the names in the file must be actual host names, not entries in the hosts file. For distributed jobs, like Glide docking, the names must be entries in the hosts file, since they are used for launching subjobs. |
||||||
|
|
Ordinarily, the hosts to use for a distributed or parallel job would be supplied using command line arguments, as explained elsewhere in this document. This variable is provided to allow the usual mechanism to be overridden when necessary. |
||||||
|
|
Specify the path name of the Maestro project into which the results of the job should be incorporated (same as entering |
||||||
|
|
Specify the path to the directory where queue log files ( |
||||||
|
|
Specify queue arguments for job submission to a queueing system. These arguments are also used when subjobs are submitted by a master job. Queue arguments specified with this environment variable are appended to those read from |
||||||
|
|
If set to 1, when the job finishes and the output files are copied back, the job directory is not removed. By default, it is removed if it did not exist already. |
||||||
|
|
If set to 1, archive the contents of the job directory and copy it back to the submission host when the job finishes. The archive appears in the output directory as jobid Equivalent to using the |
||||||
SCHRODINGER_SSH_IDENTITY
|
(Windows only) Full path to the file used for ssh connection. This can be used to communicate the location to Job Control when the file is not in the default location (%HOMEPATH%\%USERNAME%.ppk). |
||||||
|
|
Time interval for checking for stranded jobs by |
||||||
|
|
Check whether an active job is still running if its job record has not been updated for this length of time. Default: 10 minutes. |
||||||
|
|
Time interval for checking on submitted and launched jobs by |
||||||
|
|
Check whether a submitted job is still in the queue if its job record has not been updated for this length of time. Default: 5 minutes. |
||||||
|
|
Specify the full path name for the directory in which run time temporary job directories are created. Overrides the |