sge_execd.8
NAME
sge_execd, sge_loadsensor - Grid Engine job execution agent and load
sensor interface
SYNOPSIS
sge_execd [ -help ]
DESCRIPTION
sge_execd controls the Grid Engine queues local to the machine on which
sge_execd is running and executes/controls the jobs sent from
sge_qmaster(8) to be run on these queues via sge_shepherd(8) or the
shepherd_cmd of sge_conf(5).
OPTIONS
-help
Prints a listing of all options.
LOAD SENSORS
One of more load sensors may be configured for sge_execd via the
possibilities listed in sge_conf(5): the global host configuration, the
execution-host-specific cluster configuration, the default qloadsensor,
or qidle (when USE_QIDLE is set - see sge_conf(5)). The executable
path of the load sensor is invoked by sge_execd on a regular basis
(governed by load_report_time in sge_conf(5)), and delivers one or
multiple load figures for the execution host (e.g. users currently
logged in) or the complete cluster (e.g. free disk space on a network
wide scratch file system). A load sensor may be a script or a binary
executable. In either case its handling of the STDIN and STDOUT streams
and its control flow must comply with the following rules. Load
sensors are restarted if their modification time changes or they are
killed. If they read a configuration file, for instance, killing will
be necessary to pick up modifications to it unless the sensor will re-
read a modified version.
Load sensor interface
The load sensor must be written as an infinite loop, waiting at a
certain point for input from STDIN. If the string "quit" is read from
STDIN, the load sensor should exit. When an end-of-line is read from
STDIN, a load data retrieval cycle should start. The load sensor then
performs whatever operation is necessary to compute the desired load
figures. At the end of the cycle the load sensor writes the result to
stdout. The format is as follows:
o A load value report starts with a line containing only either the
word "start" or the word "begin".
o Individual load values are separated by newlines.
o Each load value report consists of three parts separated by colons
(":") and containing no blanks.
o The first part of a load value information is either the name of the
host for which load is reported or the special name "global".
o The second part is the symbolic name of the load value as defined in
the host or global complex list (see complex(5) for details). This
must be the full name of the complex, not the shortcut name. If a
load value is reported for which no entry in the host or global
complex list exists, the reported load value is not used.
o The third part is the measured load value.
o A load value report ends with a line with only the word "end".
NB. If the runtime of the language in which the sensor is written
buffers the output (e.g. Perl), ensure it is flushed on each iteration.
ENVIRONMENT VARIABLES
SGE_ROOT Specifies the location of the Grid Engine standard
configuration files.
SGE_CELL If set, specifies the default Grid Engine cell. To
address a Grid Engine cell sge_execd uses (in the order
of precedence):
The name of the cell specified in the environment
variable SGE_CELL, if it is set.
The name of the default cell, i.e. default.
SGE_DEBUG_LEVEL
If set, specifies that debug information should be
written to stderr. In addition, the level of detail in
which debug information is generated is defined.
SGE_QMASTER_PORT
If set, specifies the tcp port on which sge_qmaster(8)
is expected to listen for communication requests. Most
installations will use a services map entry for the
service "sge_qmaster" instead to define that port.
SGE_EXECD_PORT If set, specifies the tcp port on which sge_execd(8) is
expected to listen for communication requests. Most
installations will use a services map entry for the
service "sge_execd" instead to define that port.
SGE_ND If set, don't daemonize the program (for debugging).
SGE_ENABLE_COREDUMP
If set, enable core dumps on Linux when the admin_user
is not root. Linux normally disables core dumps when
the daemon has changed uid or gid. Setting
SGE_ENABLE_COREDUMP in sge_execd's environment defeats
that to enable core dumps for debugging if they are
otherwise allowed. This is typically not a big hazard
with SGE, since most information is exposed in the spool
area anyhow. Dumps will appear in the qmaster spool
directory, which need not be world-readable.
On Solaris, coreadm(1) may be used to enable such dumps.
SGE_EXECD_PIDFILE
Path name of file to which to write the daemon process
id at startup (default "execd.pid"). Note that this
must be writable by the admin user (see bootstrap(5)).
RESTRICTIONS
sge_execd usually is started from root on each machine in the Grid
Engine pool. If started by a normal user, a spool directory must be
used to which the user has read/write access. In this case only jobs
being submitted by that same user are handled correctly by the system.
FILES
<sge_root>/<cell>/common/configuration
Grid Engine global configuration
<sge_root>/<cell>/common/local_conf/<host>
Grid Engine host specific configuration
<sge_root>/<cell>/spool/<host>
Default execution host spool directory
<sge_root>/<cell>/common/act_qmaster
Grid Engine master host file
<sge_root>/bin/<arch>/qloadsensor
Default load sensor
<sge_root>/bin/<arch>/qidle
Idle load sensor per USE_QIDLE in execd_params
<sge_root>/<cell>/common/sgepasswd
Password information used on Microsoft Windows hosts. See
sgepasswd(5).
SEE ALSO
sge_intro(1), sge_conf(5), complex(5), load_parameters(5),
sgepasswd(5), sge_shepherd(8), sge_qmaster(8).
COPYRIGHT
See sge_intro(1) for a full statement of rights and permissions.
SGE 8.1.3pre 2011-12-29 SGE_EXECD(8)
Man(1) output converted with
man2html