diff --git a/NEWS b/NEWS
index 0e560ddffac7f1449909952f738af6a8051e035a..e172b22247f9b89ca8729ee4c8c1395bb0d161e9 100644
--- a/NEWS
+++ b/NEWS
@@ -26,7 +26,8 @@ documents those changes that are of interest to users and admins.
* Changes in SLURM 2.0.1
========================
-- Fix, truncate time of start and end for job steps in sacct.
- -- initialize all messages to slurmdbd
+ -- Initialize all messages to slurmdbd. Previously uninitialized string could
+ cause slurmctld to fail with invalid memory reference.
-- BLUEGENE - Fix, for when trying to finish a torus on a block already
visited. Even though this may be possible electrically this isn't valid
in the under lying infrastructure.
@@ -45,6 +46,11 @@ documents those changes that are of interest to users and admins.
-- Fix "-Q" (quiet) option for salloc and sbatch which was previously
ignored.
-- BLUEGENE - fix for finding odd shaped blocks in dynamic mode.
+ -- Fix logic supporting SuspendRate and ResumeRate configuration parameters.
+ Previous logic was changing state of one too many nodes per minute.
+ -- Save new reservation state file on shutdown (even if no changes).
+ -- Fix, when partitions are deleted the sched and select plugins are notified.
+ -- Fix linking problem that prevented checkpoint/aix from working.
* Changes in SLURM 2.0.0
========================
diff --git a/doc/html/power_save.shtml b/doc/html/power_save.shtml
index aa29216ce51de33b8920c000b0bd79462683ba04..0254e701892a4d1f537f04454f9e998e2252226e 100644
--- a/doc/html/power_save.shtml
+++ b/doc/html/power_save.shtml
@@ -43,6 +43,8 @@ Maximum number of nodes to be placed into power saving mode
per minute.
A value of zero results in no limits being imposed.
The default value is 60.
+Rate calculations are performed using integers, so rounding errors
+may be significant for very small values (1 or 2 nodes per minute).
Use this to prevent rapid drops in power requirements.</li>
<li><b>ResumeRate</b>:
@@ -50,6 +52,8 @@ Maximum number of nodes to be removed from power saving mode
per minute.
A value of zero results in no limits being imposed.
The default value is 300.
+Rate calculations are performed using integers, so rounding errors
+may be significant for very small values (1 or 2 nodes per minute).
Use this to prevent rapid increases in power requirements.</li>
<li><b>SuspendProgram</b>:
@@ -219,6 +223,6 @@ In order to minimize this risk, when the <i>slurmctld</i> daemon is
started and node which should be allocated to a job fails to respond,
the <b>ResumeProgram</b> will be executed (possibly for a second time).</p>
-<p style="text-align:center;">Last modified 27 May 2009</p>
+<p style="text-align:center;">Last modified 2 June 2009</p>
<!--#include virtual="footer.txt"-->
diff --git a/doc/man/man5/slurm.conf.5 b/doc/man/man5/slurm.conf.5
index 78ddbb16e462c74288c5739519d43867d2c2b55c..e1d931226ef717fdf84f3663464c3c3ad791c11e 100644
--- a/doc/man/man5/slurm.conf.5
+++ b/doc/man/man5/slurm.conf.5
@@ -1018,8 +1018,10 @@ operation by \fBResumeProgram\fR.
The value is number of nodes per minute and it can be used to prevent
power surges if a large number of nodes in power save mode are
assigned work at the same time (e.g. a large job starts).
-A value of zero results in no limits being imposed.
+A value of zero results in no limits being imposed.
The default value is 300 nodes per minute.
+Rate calculations are performed using integers, so rounding errors
+may be significant for very small values (1 or 2 nodes per minute).
Related configuration options include \fBResumeTimeout\fR, \fBResumeProgram\fR,
\fBSuspendRate\fR, \fBSuspendTime\fR, \fBResumeTimeout\fR, \fBSuspendProgram\fR,
\fBSuspendExcNodes\fR, and \fBSuspendExcParts\fR.
@@ -1441,6 +1443,8 @@ The value is number of nodes per minute and it can be used to prevent
a large drop in power power consumption (e.g. after a large job completes).
A value of zero results in no limits being imposed.
The default value is 60 nodes per minute.
+Rate calculations are performed using integers, so rounding errors
+may be significant for very small values (1 or 2 nodes per minute).
Related configuration options include \fBResumeTimeout\fR, \fBResumeProgram\fR,
\fBResumeRate\fR, \fBSuspendProgram\fR, \fBSuspendTime\fR, \fBSuspendTimeout\fR,
\fBSuspendExcNodes\fR, and \fBSuspendExcParts\fR.
diff --git a/src/plugins/checkpoint/aix/checkpoint_aix.c b/src/plugins/checkpoint/aix/checkpoint_aix.c
index 39a18402515232cd1a34bd4572c43a447098a061..ee4a374da7435b639e0045f7d467f91ad588d0c4 100644
--- a/src/plugins/checkpoint/aix/checkpoint_aix.c
+++ b/src/plugins/checkpoint/aix/checkpoint_aix.c
@@ -58,6 +58,7 @@
#include <slurm/slurm.h>
#include <slurm/slurm_errno.h>
+#include "src/common/slurm_xlator.h"
#include "src/common/list.h"
#include "src/common/log.h"
#include "src/common/pack.h"
@@ -67,6 +68,13 @@
#include "src/slurmctld/agent.h"
#include "src/slurmctld/slurmctld.h"
+/* These are defined here so when we link with something other than
+ * the slurmctld we will have these symbols defined. They will get
+ * overwritten when linking with the slurmctld.
+ */
+struct node_record *node_record_table_ptr = NULL;
+int node_record_count = 0;
+
struct check_job_info {
uint16_t disabled; /* counter, checkpointable only if zero */
uint16_t node_cnt;
diff --git a/src/plugins/select/bluegene/plugin/block_sys.c b/src/plugins/select/bluegene/plugin/block_sys.c
index 09beaee2ad1118b3494c5b68af86a7ee983bde43..3053da497e29719cb40bc90a98fcf3a4b73d0869 100755
--- a/src/plugins/select/bluegene/plugin/block_sys.c
+++ b/src/plugins/select/bluegene/plugin/block_sys.c
@@ -1155,7 +1155,7 @@ extern int load_state_file(List curr_block_list, char *dir_name)
bg_record->state = bg_info_record->state;
bg_record->job_running = NO_JOB_RUNNING;
- bg_record->bp_count = bit_size(node_bitmap);
+ bg_record->bp_count = bit_set_count(node_bitmap);
bg_record->node_cnt = bg_info_record->node_cnt;
if(bg_conf->bp_node_cnt > bg_record->node_cnt) {
ionodes = bg_conf->bp_node_cnt
diff --git a/src/slurmctld/agent.c b/src/slurmctld/agent.c
index 6f235cb59fab40ae31e5da264e1c7ade1009d60f..6775b2a0945bcd70dc391c8e9172ef75aa672a0f 100644
--- a/src/slurmctld/agent.c
+++ b/src/slurmctld/agent.c
@@ -1520,7 +1520,8 @@ static int _batch_launch_defer(queued_request_t *queued_req_ptr)
if (!IS_NODE_POWER_SAVE(node_ptr) && !IS_NODE_NO_RESPOND(node_ptr)) {
/* ready to launch, adjust time limit for boot time */
int resume_timeout = slurm_get_resume_timeout();
- if ((job_ptr->start_time + resume_timeout) >= now)
+ if ((job_ptr->time_limit != INFINITE)
+ && ((job_ptr->start_time + resume_timeout) >= now))
job_ptr->end_time = now + (job_ptr->time_limit * 60);
queued_req_ptr->last_attempt = (time_t) 0;
return 0;
diff --git a/src/slurmctld/controller.c b/src/slurmctld/controller.c
index 300ffe48f4f1d2ba5e31ef19810dd22434b90d5b..8fa161a709c5746ddc3a68f3adeb83bb5a429275 100644
--- a/src/slurmctld/controller.c
+++ b/src/slurmctld/controller.c
@@ -1358,8 +1358,9 @@ void save_all_state(void)
{
/* Each of these functions lock their own databases */
schedule_job_save();
- schedule_part_save();
schedule_node_save();
+ schedule_part_save();
+ schedule_resv_save();
schedule_trigger_save();
select_g_state_save(slurmctld_conf.state_save_location);
dump_assoc_mgr_state(slurmctld_conf.state_save_location);
diff --git a/src/slurmctld/job_scheduler.c b/src/slurmctld/job_scheduler.c
index 361d9af918d1d73dc3270971f79141ed9d92e82c..1557098b21f50eec1810893429d6f886c67daac4 100644
--- a/src/slurmctld/job_scheduler.c
+++ b/src/slurmctld/job_scheduler.c
@@ -393,7 +393,8 @@ extern int schedule(void)
bit_not(job_ptr->part_ptr->node_bitmap);
}
} else if (error_code == ESLURM_RESERVATION_NOT_USABLE) {
- if (job_ptr->resv_ptr && job_ptr->resv_ptr->node_bitmap) {
+ if (job_ptr->resv_ptr
+ && job_ptr->resv_ptr->node_bitmap) {
bit_not(job_ptr->resv_ptr->node_bitmap);
bit_and(avail_node_bitmap,
job_ptr->resv_ptr->node_bitmap);
diff --git a/src/slurmctld/node_scheduler.c b/src/slurmctld/node_scheduler.c
index 83c1f52159c7fdc9673f614a3567025bd70cf3d2..479f9da689396e5ad2376735097ba0b7eaa78555 100644
--- a/src/slurmctld/node_scheduler.c
+++ b/src/slurmctld/node_scheduler.c
@@ -1095,8 +1095,8 @@ extern int select_nodes(struct job_record *job_ptr, bool test_only,
(365 * 24 * 60 * 60); /* secs in year */
else
job_ptr->end_time = job_ptr->start_time +
- (job_ptr->time_limit * 60); /* secs */
-
+ (job_ptr->time_limit * 60); /* secs */
+
if (select_g_job_begin(job_ptr) != SLURM_SUCCESS) {
/* Leave job queued, something is hosed */
error("select_g_job_begin(%u): %m", job_ptr->job_id);
diff --git a/src/slurmctld/partition_mgr.c b/src/slurmctld/partition_mgr.c
index 14417cbbee946dd4ff3b7b48a3db0ef3a18a43c5..53d2dd38b3b1ac4b32660671df536c192ab25aa5 100644
--- a/src/slurmctld/partition_mgr.c
+++ b/src/slurmctld/partition_mgr.c
@@ -880,7 +880,8 @@ extern int update_part (update_part_msg_t * part_desc, bool create_flag)
info("update_part: DefaultTime would exceed MaxTime for "
"partition %s", part_desc->name);
} else if (part_desc->default_time != NO_VAL) {
- info("update_part: setting default_time to %u for partition %s",
+ info("update_part: setting default_time to %u "
+ "for partition %s",
part_desc->default_time, part_desc->name);
part_ptr->default_time = part_desc->default_time;
}
@@ -963,7 +964,8 @@ extern int update_part (update_part_msg_t * part_desc, bool create_flag)
info("update_part: setting default partition to %s",
part_desc->name);
} else if (strcmp(default_part_name, part_desc->name) != 0) {
- info("update_part: changing default partition from %s to %s",
+ info("update_part: changing default "
+ "partition from %s to %s",
default_part_name, part_desc->name);
}
xfree(default_part_name);
@@ -1027,7 +1029,8 @@ extern int update_part (update_part_msg_t * part_desc, bool create_flag)
xfree(part_ptr->nodes);
part_ptr->nodes = backup_node_list;
} else {
- info("update_part: setting nodes to %s for partition %s",
+ info("update_part: setting nodes to %s "
+ "for partition %s",
part_ptr->nodes, part_desc->name);
xfree(backup_node_list);
}
@@ -1328,5 +1331,9 @@ extern int delete_partition(delete_part_msg_t *part_desc_ptr)
list_delete_all(part_list, list_find_part, part_desc_ptr->name);
last_part_update = time(NULL);
+ slurm_sched_partition_change(); /* notify sched plugin */
+ select_g_reconfigure(); /* notify select plugin too */
+ reset_job_priority(); /* free jobs */
+
return SLURM_SUCCESS;
}
diff --git a/src/slurmctld/power_save.c b/src/slurmctld/power_save.c
index c1c88f20c786778025a0ed12ec7742bf5dc2d3fb..4c94c176b6d20fec20bd47062609841cc0b160f4 100644
--- a/src/slurmctld/power_save.c
+++ b/src/slurmctld/power_save.c
@@ -139,7 +139,7 @@ static void _do_power_work(void)
if (susp_state)
susp_total++;
if (susp_state &&
- ((suspend_rate == 0) || (suspend_cnt <= suspend_rate)) &&
+ ((suspend_rate == 0) || (suspend_cnt < suspend_rate)) &&
(bit_test(suspend_node_bitmap, i) == 0) &&
(IS_NODE_ALLOCATED(node_ptr) ||
(node_ptr->last_idle > (now - idle_time)))) {
@@ -157,7 +157,7 @@ static void _do_power_work(void)
}
if (run_suspend &&
(susp_state == 0) &&
- ((resume_rate == 0) || (resume_cnt <= resume_rate)) &&
+ ((resume_rate == 0) || (resume_cnt < resume_rate)) &&
IS_NODE_IDLE(node_ptr) &&
(!IS_NODE_COMPLETING(node_ptr)) &&
(node_ptr->last_idle < (now - idle_time)) &&
diff --git a/src/slurmd/slurmd/req.c b/src/slurmd/slurmd/req.c
index d864f9d14cb295b07ff4ce322674e6fe7fd89317..d9d2cd5fa4864483280976eb3794ee65ab9e6c34 100644
--- a/src/slurmd/slurmd/req.c
+++ b/src/slurmd/slurmd/req.c
@@ -3173,6 +3173,8 @@ _run_prolog(uint32_t jobid, uid_t uid, char *resv_id,
char *my_prolog;
char **my_env = _build_env(jobid, uid, resv_id, spank_job_env,
spank_job_env_size);
+ time_t start_time = time(NULL), diff_time;
+ static uint16_t msg_timeout = 0;
slurm_mutex_lock(&conf->config_mutex);
my_prolog = xstrdup(conf->prolog);
@@ -3182,6 +3184,14 @@ _run_prolog(uint32_t jobid, uid_t uid, char *resv_id,
xfree(my_prolog);
_destroy_env(my_env);
+ diff_time = difftime(time(NULL), start_time);
+ if (msg_timeout == 0)
+ msg_timeout = slurm_get_msg_timeout();
+ if (diff_time >= msg_timeout) {
+ error("prolog for job %u ran for %d seconds",
+ jobid, diff_time);
+ }
+
return error_code;
}