diff --git a/doc/man/man1/slaunch.1 b/doc/man/man1/slaunch.1
index 0b51bd9207af4c660c8d16f44690a5bc916b66b5..4474b5477e616d107974d6a1f006e6a0e3381444 100644
--- a/doc/man/man1/slaunch.1
+++ b/doc/man/man1/slaunch.1
@@ -15,39 +15,36 @@ slaunch launches a parallel application (a \fBjob step\fR in SLURM parlance) on
\fB\-\-jobid\fR <\fIJOBID\fP>
The job allocation under which the parallel application should be launched.
.TP
-\fB\-n\fR, \fB\-\-ntasks\fR=<\fInumber\fR>
+\fB\-n\fR, \fB\-\-ntasks\fR[=]<\fInumber\fR>
Specify the number of processes to launch. The default is one process per node, but note that the \-c parameter will change this default.
.TP
-\fB\-N\fR, \fB\-\-nodes\fR=<\fInumber\fR>
+\fB\-N\fR, \fB\-\-nodes\fR[=]<\fInumber\fR>
Specify the number of nodes to be used by this job step. By default,
slaunch will use all of the nodes in the specified job allocation.
.TP
-\fB\-r\fR, \fB\-\-relative\fR=<\fInumber\fR>
+\fB\-r\fR, \fB\-\-relative\fR[=]<\fInumber\fR>
Specify the first node in the allocation on which this job step will be launched. Counting starts at zero, thus the first node in the job allocation is node 0. The default value for \-\-relative is zero.
.TP
-\fB\-c\fR, \fB\-\-cpus\-per\-task\fR=<\fIncpus\fR>
+\fB\-c\fR, \fB\-\-cpus\-per\-task\fR[=]<\fIncpus\fR>
Specify that each task requires \fIncpus\fR number of CPUs. Useful for applications in which each task will launch multiple threads and can therefore benefit from there being free processors on the node.
+.TP
+\fB\-w\fR, \fB\-\-nodelist\fR[=]<\fInodelist\fR>
+Request a specific list of node names. The list may be specified as a comma\-separated list of node names, or a range of node names (e.g. mynode[1\-5,7,...]).
+
+If the configured SLURM switch plugin supports the "arbitrary" method of task distribution, the nodelist may also contain nodenames more than once, and in any order. Each node will then represent one task in the job step. For example, given a nodelist "mynode[2,8,2,2,8]" tasks 0, 2, and 3 will run on mynode2, and tasks 1 and 4 will run on mynode8.
.TP
\fB\-\-nice\fR=<\fIadjustment\fR>
Run the job with an adjusted scheduling priority. With no adjustment value the scheduling priority is decreased by 100. The adjustment range is from \-10000 (highest priority) to 10000 (lowest priority). Only privileged users can specify a negative adjustment. NOTE: This option is presently ignored if SchedulerType=sched/maui.
-.PP
-Help options
-.TP
-\fB\-\-help\fR
-Display verbose help message and exit.
-.TP
-\fB\-\-usage\fR
-Display brief help message and exit.
-.PP
-Other options
-.TP
+.TP
+\fB\-\-help\fR
+Output help information and exit.
+.TP
\fB\-V\fR, \fB\-\-version\fR
-Display version information and exit.
-
+Output version information and exit.
.SH "ENVIRONMENT VARIABLES"
.LP
.TP
diff --git a/src/slaunch/opt.c b/src/slaunch/opt.c
index 977c92fbc54f6cee76b56f384230d2b34ea6f923..868264aa6520cb56edd9cb37c8a649ea3b8b4e97 100644
--- a/src/slaunch/opt.c
+++ b/src/slaunch/opt.c
@@ -71,6 +71,7 @@
#include "src/common/plugstack.h"
#include "src/common/optz.h"
#include "src/common/read_config.h" /* getnodename() */
+#include "src/common/hostlist.h"
#include "src/slaunch/attach.h"
#include "src/common/mpi.h"
@@ -645,6 +646,7 @@ static void _opt_default()
opt.jobid_set = false;
opt.distribution = SLURM_DIST_CYCLIC;
+ opt.distribution_set = false;
opt.local_ofname = NULL;
opt.local_ifname = NULL;
@@ -663,9 +665,6 @@ static void _opt_default()
opt.max_wait = slurm_get_wait_time();
- opt.quit_on_intr = false;
- opt.disable_status = false;
-
opt.quiet = 0;
_verbose = 0;
opt.slurmd_debug = LOG_LEVEL_QUIET;
@@ -679,7 +678,6 @@ static void _opt_default()
opt.contiguous = false;
opt.exclusive = false;
opt.nodelist = NULL;
- opt.exc_nodes = NULL;
opt.max_launch_time = 120;/* 120 seconds to launch job */
opt.max_exit_timeout= 60; /* Warn user 60 seconds after task exit */
opt.msg_timeout = 5; /* Default launch msg timeout */
@@ -756,7 +754,6 @@ env_vars_t env_vars[] = {
{"SLAUNCH_STDOUTMODE", OPT_STRING, &opt.local_ofname, NULL },
{"SLAUNCH_TIMELIMIT", OPT_INT, &opt.time_limit, NULL },
{"SLAUNCH_WAIT", OPT_INT, &opt.max_wait, NULL },
- {"SLAUNCH_DISABLE_STATUS",OPT_INT, &opt.disable_status,NULL },
{"SLAUNCH_MPI_TYPE", OPT_MPI, NULL, NULL },
{"SLAUNCH_SRUN_COMM_IFHN",OPT_STRING, &opt.ctrl_comm_ifhn,NULL },
{"SLAUNCH_SRUN_MULTI", OPT_MULTI, NULL, NULL },
@@ -820,8 +817,10 @@ _process_env_var(env_vars_t *e, const char *val)
if (dt == -1) {
error("\"%s=%s\" -- invalid distribution type. "
"ignoring...", e->var, val);
- } else
+ } else {
opt.distribution = dt;
+ opt.distribution_set = true;
+ }
break;
case OPT_CPU_BIND:
@@ -921,7 +920,6 @@ void set_options(const int argc, char **argv, int first)
{"local-output", required_argument, 0, 'o'},
{"remote-output", required_argument, 0, 'O'},
{"overcommit", no_argument, 0, 'C'},
- {"quit-on-interrupt", no_argument, 0, 'q'},
{"quiet", no_argument, 0, 'Q'},
{"relative", required_argument, 0, 'r'},
{"no-rotate", no_argument, 0, 'R'},
@@ -931,8 +929,6 @@ void set_options(const int argc, char **argv, int first)
{"version", no_argument, 0, 'V'},
{"nodelist", required_argument, 0, 'w'},
{"wait", required_argument, 0, 'W'},
- {"exclude", required_argument, 0, 'x'},
- {"disable-status", no_argument, 0, 'X'},
{"no-allocate", no_argument, 0, 'Z'},
{"contiguous", no_argument, 0, LONG_OPT_CONT},
{"exclusive", no_argument, 0, LONG_OPT_EXCLUSIVE},
@@ -964,7 +960,7 @@ void set_options(const int argc, char **argv, int first)
{NULL, 0, 0, 0}
};
char *opt_string = "+c:Cd:D:e:E:g:i:I:J:kKlm:n:N:"
- "o:O:qQr:R:t:uvVw:W:x:XZ";
+ "o:O:Qr:R:t:uvVw:W:Z";
struct option *optz = spank_option_table_create (long_options);
@@ -1078,6 +1074,8 @@ void set_options(const int argc, char **argv, int first)
error("distribution type `%s' "
"is not recognized", optarg);
exit(1);
+ } else {
+ opt.distribution_set = true;
}
break;
case (int)'n':
@@ -1114,9 +1112,6 @@ void set_options(const int argc, char **argv, int first)
else
opt.remote_ofname = xstrdup(optarg);
break;
- case (int)'q':
- opt.quit_on_intr = true;
- break;
case (int) 'Q':
if(!first && opt.quiet)
break;
@@ -1170,15 +1165,6 @@ void set_options(const int argc, char **argv, int first)
case (int)'W':
opt.max_wait = _get_int(optarg, "wait");
break;
- case (int)'x':
- xfree(opt.exc_nodes);
- opt.exc_nodes = xstrdup(optarg);
- if (!_valid_node_list(&opt.exc_nodes))
- exit(1);
- break;
- case (int)'X':
- opt.disable_status = true;
- break;
case (int)'Z':
opt.no_alloc = true;
uname(&name);
@@ -1417,12 +1403,6 @@ static void _opt_args(int argc, char **argv)
}
}
- if (!opt.num_nodes_set
- && opt.num_tasks_set && opt.num_tasks < opt.num_nodes)
- opt.num_nodes = opt.num_tasks;
- if (!opt.num_tasks_set)
- opt.num_tasks = opt.num_nodes;
-
if (!_opt_verify())
exit(1);
}
@@ -1434,6 +1414,66 @@ static void _opt_args(int argc, char **argv)
static bool _opt_verify(void)
{
bool verified = true;
+ hostlist_t hl = NULL;
+ hostlist_t hl_unique = NULL;
+
+ if (opt.nodelist != NULL) {
+ hl = hostlist_create(opt.nodelist);
+ hl_unique = hostlist_copy(hl);
+ hostlist_uniq(hl_unique);
+ }
+
+ /* Automatically trigger arbitrary task distribution mode if a hostlist
+ was specified, and the sorted unique hostlist does not match the
+ unsorted hostlist */
+ if (!opt.distribution_set && opt.nodelist != NULL) {
+ if (hostlist_count(hl) != hostlist_count(hl_unique)) {
+ opt.distribution = SLURM_DIST_ARBITRARY;
+ } else {
+ /* They are the same length, but are they in the
+ same order? There is an easy way to tell;
+ turn them both back into a strings and compare. */
+ char buf1[8192], buf2[8192];
+ buf1[0] = buf2[0] = '\0';
+ hostlist_ranged_string(hl, 8192, buf1);
+ hostlist_ranged_string(hl_unique, 8192, buf2);
+ if (strcmp(buf1, buf2) != 0) {
+ opt.distribution = SLURM_DIST_ARBITRARY;
+ }
+ }
+ }
+
+ if (!opt.num_nodes_set && opt.nodelist != NULL) {
+ /* Need to ask for at as many nodes as are in the
+ nodelist, otherwise the controller gets confused. */
+ opt.num_nodes = hostlist_count(hl_unique);
+ }
+ if (!opt.num_nodes_set
+ && opt.num_tasks_set && opt.num_tasks < opt.num_nodes)
+ opt.num_nodes = opt.num_tasks;
+ if (opt.num_nodes_set
+ && opt.num_nodes != hostlist_count(hl_unique)) {
+ if (opt.num_nodes > hostlist_count(hl_unique)) {
+ error("Asked for more nodes (%d) "
+ "than listed in the nodelist (%d)",
+ opt.num_nodes, hostlist_count(hl_unique));
+ verified = false;
+ } else { /* num_nodes < hostlist_count */
+ /* shrink the nodelist instead of an error? */
+ error("Asked for fewer nodes (%d) "
+ "than listed in the nodelist (%d)",
+ opt.num_nodes, hostlist_count(hl_unique));
+ verified = false;
+ }
+ }
+ if (!opt.num_tasks_set) {
+ if (opt.nodelist != NULL) {
+ opt.num_tasks = hostlist_count(hl);
+ } else {
+ opt.num_tasks = opt.num_nodes;
+ }
+ }
+
if (!opt.jobid_set) {
error("A job ID MUST be specified on the command line,");
@@ -1461,19 +1501,13 @@ static bool _opt_verify(void)
verified = false;
}
- if (opt.no_alloc && opt.exc_nodes) {
- error("can not specify --exclude list with -Z, --no-allocate.");
- verified = false;
- }
-
if (opt.no_alloc && opt.relative) {
error("do not specify -r,--relative with -Z,--no-allocate.");
verified = false;
}
- if (opt.relative && (opt.exc_nodes || opt.nodelist)) {
- error("-r,--relative not allowed with "
- "-w,--nodelist or -x,--exclude.");
+ if (opt.relative && opt.nodelist) {
+ error("-r,--relative not allowed with -w,--nodelist.");
verified = false;
}
@@ -1510,20 +1544,6 @@ static bool _opt_verify(void)
core_format_enable (opt.core_type);
- /* massage the numbers */
- if (opt.num_nodes_set && !opt.num_tasks_set) {
- /* 1 proc / node default */
- opt.num_tasks = opt.num_nodes;
- } else if (opt.num_nodes_set && opt.num_tasks_set) {
- if (opt.num_tasks < opt.num_nodes) {
- info ("Warning: can't run %d tasks on %d "
- "nodes, setting nnodes to %d",
- opt.num_tasks, opt.num_nodes, opt.num_tasks);
- opt.num_nodes = opt.num_tasks;
- }
-
- } /* else if (opt.num_tasks_set && !opt.num_nodes_set) */
-
if (opt.labelio && opt.unbuffered) {
error("Do not specify both -l (--label) and "
"-u (--unbuffered)");
@@ -1553,6 +1573,8 @@ static bool _opt_verify(void)
verified = false;
}
+ hostlist_destroy(hl);
+ hostlist_destroy(hl_unique);
return verified;
}
@@ -1659,9 +1681,6 @@ static char *print_constraints()
if (opt.nodelist != NULL)
xstrfmtcat(buf, "nodelist=%s ", opt.nodelist);
- if (opt.exc_nodes != NULL)
- xstrfmtcat(buf, "exclude=%s ", opt.exc_nodes);
-
if (opt.constraints != NULL)
xstrfmtcat(buf, "constraints=`%s' ", opt.constraints);
@@ -1723,7 +1742,9 @@ static void _opt_list()
info("jobid : %u %s", opt.jobid,
opt.jobid_set ? "(set)" : "(default)");
info("job name : \"%s\"", opt.job_name);
- info("distribution : %s", format_task_dist_states(opt.distribution));
+ info("distribution : %s %s",
+ format_task_dist_states(opt.distribution),
+ opt.distribution_set ? "(set)" : "(default)");
info("cpu_bind : %s",
opt.cpu_bind == NULL ? "default" : opt.cpu_bind);
info("mem_bind : %s",
@@ -1788,7 +1809,7 @@ static void _usage(void)
" [--prolog=fname] [--epilog=fname]\n"
" [--task-prolog=fname] [--task-epilog=fname]\n"
" [--ctrl-comm-ifhn=addr] [--multi-prog]\n"
-" [-w hosts...] [-x hosts...] executable [args...]\n");
+" [-w hosts...] executable [args...]\n");
}
static void _help(void)
@@ -1823,8 +1844,6 @@ static void _help(void)
" -b, --batch submit as batch job for later execution\n"
" -W, --wait=sec seconds to wait after first task exits\n"
" before killing job\n"
-" -q, --quit-on-interrupt quit on single Ctrl-C\n"
-" -X, --disable-status Disable Ctrl-C status feature\n"
" -v, --verbose verbose mode (multiple -v's increase verbosity)\n"
" -Q, --quiet quiet mode (suppress informational messages)\n"
" -d, --slurmd-debug=level slurmd debug level\n"
@@ -1847,7 +1866,6 @@ static void _help(void)
" --contiguous demand a contiguous range of nodes\n"
" -C, --constraint=list specify a list of constraints\n"
" -w, --nodelist=hosts... request a specific list of hosts\n"
-" -x, --exclude=hosts... exclude a specific list of hosts\n"
" -Z, --no-allocate don't allocate nodes (must supply -w)\n"
"\n"
"Consumable resources related options:\n"
diff --git a/src/slaunch/opt.h b/src/slaunch/opt.h
index 4c2afa1df35613fdec3e9ae85f25ef56926db01a..1dd35f35bbfb4cdaeecbb85bf38b55e337f18a11 100644
--- a/src/slaunch/opt.h
+++ b/src/slaunch/opt.h
@@ -80,6 +80,7 @@ typedef struct slaunch_options {
int time_limit; /* --time, -t */
enum task_dist_states
distribution; /* --distribution=, -m dist */
+ bool distribution_set;
char *job_name; /* --job-name=, -J name */
unsigned int jobid; /* --jobid=jobid */
bool jobid_set; /* true of jobid explicitly set */
@@ -101,8 +102,6 @@ typedef struct slaunch_options {
bool no_kill; /* --no-kill, -k */
bool kill_bad_exit; /* --kill-on-bad-exit, -K */
int max_wait; /* --wait, -W */
- bool quit_on_intr; /* --quit-on-interrupt, -q */
- bool disable_status; /* --disable-status, -X */
int quiet;
bool parallel_debug; /* slaunch controlled by debugger */
bool debugger_test; /* --debugger-test */
@@ -117,7 +116,6 @@ typedef struct slaunch_options {
char *constraints; /* --constraints=, -C constraint*/
bool contiguous; /* --contiguous */
char *nodelist; /* --nodelist=node1,node2,... */
- char *exc_nodes; /* --exclude=node1,node2,... -x */
int relative; /* --relative -r N */
bool relative_set; /* true if --relative set explicitly */
bool no_alloc; /* --no-allocate, -Z */
diff --git a/src/slaunch/slaunch.c b/src/slaunch/slaunch.c
index d5575f0c8abf1cfc18d8073237b81f27bff84eb7..816637581406a0508465156d9e5c8ca2b92a7a23 100644
--- a/src/slaunch/slaunch.c
+++ b/src/slaunch/slaunch.c
@@ -151,18 +151,15 @@ int slaunch(int argc, char **argv)
xstrfmtcat(totalview_jobid, "%u", step_req.job_id);
step_req.user_id = getuid();
step_req.node_count = opt.num_nodes;
- if (opt.num_tasks_set)
- step_req.num_tasks = opt.num_tasks;
- else
- step_req.num_tasks = opt.num_nodes;
+ step_req.num_tasks = opt.num_tasks;
/* step_req.cpu_count = opt.cpus_per_task; */
- step_req.cpu_count = 0;
+ step_req.cpu_count = 0; /* FIXME */
step_req.relative = opt.relative;
- step_req.task_dist = SLURM_DIST_CYCLIC;
+ step_req.task_dist = opt.distribution;
step_req.port = 0; /* historical, used by srun */
step_req.host = NULL; /* historical, used by srun */
- step_req.node_list = NULL;
- step_req.network = NULL;
+ step_req.node_list = opt.nodelist;
+ step_req.network = opt.network;
step_req.name = opt.job_name;
step_ctx = slurm_step_ctx_create(&step_req);