diff --git a/NEWS b/NEWS
index 7fbfa2f1a9aec83c887fab6286325c3202dd70f2..decc6ac58249f27fd223b7cc00dd78c5b8bd8a26 100644
--- a/NEWS
+++ b/NEWS
@@ -79,6 +79,13 @@ documents those changes that are of interest to users and administrators.
diversion.
-- Pending job array records will be combined into single line by default,
even if started and requeued or modified.
+ -- Fix sacct --format=nnodes to print out correct information for pending
+ jobs.
+ -- Make is so 'scontrol update job 1234 qos='' will set the qos back to
+ the default qos for the association.
+ -- Add [Alloc|Req]Nodes to sacct to be more like cpus.
+ -- Fix sacct documentation about [Alloc|Req]TRES
+ -- Put node count in TRES string for steps.
* Changes in Slurm 15.08.0
==========================
@@ -450,6 +457,7 @@ documents those changes that are of interest to users and administrators.
submitted.
-- Correct counting for job array limits, job count limit underflow possible
when master cancellation of master job record.
+ -- For pending jobs have sacct print 0 for nnodes instead of the bogus 2.
* Changes in Slurm 14.11.9
==========================
diff --git a/doc/man/man1/sacct.1 b/doc/man/man1/sacct.1
index 5dfa2502f40f4a0ea2fcc2b51eca1e9871362f85..041cc37205640fb7d41c3c3ee5b51630682ca117 100644
--- a/doc/man/man1/sacct.1
+++ b/doc/man/man1/sacct.1
@@ -126,25 +126,28 @@ Print a list of fields that can be specified with the \f3\-\-format\fP option.
.ft 3
Fields available:
-AllocCPUS Account AssocID AveCPU
-AveCPUFreq AveDiskRead AveDiskWrite AvePages
-AveRSS AveVMSize BlockID Cluster
-Comment ConsumedEnergy CPUTime CPUTimeRAW
-DerivedExitCode Elapsed Eligible End
-ExitCode GID Group JobID
-JobIDRaw JobName Layout MaxDiskRead
-MaxDiskReadNode MaxDiskReadTask MaxDiskWrite MaxDiskWriteNode
-MaxDiskWriteTask MaxPages MaxPagesNode MaxPagesTask
-MaxRSS MaxRSSNode MaxRSSTask MaxVMSize
-MaxVMSizeNode MaxVMSizeTask MinCPU MinCPUNode
-MinCPUTask NCPUS NNodes NodeList
-NTasks Priority Partition QOSRAW
-ReqCPUFreqMin ReqCPUFreqMax ReqCPUFreqGov ReqCPUs
-ReqMem Reservation ReservationId Reserved
-ResvCPU ResvCPURAW Start State
-Submit Suspended SystemCPU Timelimit
-TotalCPU Tres UID User
-UserCPU WCKey WCKeyID
+AllocCPUS AllocGRES AllocNodes AllocTRES
+Account AssocID AveCPU AveCPUFreq
+AveDiskRead AveDiskWrite AvePages AveRSS
+AveVMSize BlockID Cluster Comment
+ConsumedEnergy ConsumedEnergyRaw CPUTime CPUTimeRAW
+DerivedExitCode Elapsed Eligible End
+ExitCode GID Group JobID
+JobIDRaw JobName Layout MaxDiskRead
+MaxDiskReadNode MaxDiskReadTask MaxDiskWrite MaxDiskWriteNode
+MaxDiskWriteTask MaxPages MaxPagesNode MaxPagesTask
+MaxRSS MaxRSSNode MaxRSSTask MaxVMSize
+MaxVMSizeNode MaxVMSizeTask MinCPU MinCPUNode
+MinCPUTask NCPUS NNodes NodeList
+NTasks Priority Partition QOS
+QOSRAW ReqCPUFreq ReqCPUFreqMin ReqCPUFreqMax
+ReqCPUFreqGov ReqCPUS ReqGRES ReqMem
+ReqNodes ReqTRES Reservation ReservationId
+Reserved ResvCPU ResvCPURAW Start
+State Submit Suspended SystemCPU
+Timelimit TotalCPU TRESAlloc TRESReq
+UID User UserCPU WCKey
+WCKeyID
.ft 1
.fi
@@ -441,6 +444,16 @@ Count of allocated CPUs. Equivalant to \f3NCPUS\fP.
\f3AllocGRES\fP
Names and counts of generic resources allocated.
+.TP
+\f3AllocNodes\fP
+Number of nodes allocated to the job/step. 0 if the job is pending.
+
+.TP
+\f3AllocTres\fP
+Trackable resources. These are the resources allocated to the job/step
+after the job started running. For pending jobs this should be blank.
+For more details see AccountingStorageTRES in slurm.conf.
+
.TP
\f3Account\fP
Account the job ran under.
@@ -677,7 +690,8 @@ List of nodes in job/step.
.TP
\f3NNodes\fP
-Number of nodes in a job or step.
+Number of nodes in a job or step. If the job is running, or ran, this count
+will be the number allocated, else the number will be the number requested.
.TP
\f3NTasks\fP
@@ -718,6 +732,17 @@ Minimum required memory for the job, in MB. A 'c' at the end of
number represents Memory Per CPU, a 'n' represents Memory Per Node.
Note: This value is only from the job allocation, not the step.
+.TP
+\f3ReqNodes\fP
+Requested minimum Node count for the job/step.
+
+.TP
+\f3ReqTres\fP
+Trackable resources. These are the minimum resource counts requested by the
+job/step at submission time.
+For more details see AccountingStorageTRES in slurm.conf.
+
+
.TP
\f3Reservation\fP
Reservation Name.
@@ -791,11 +816,6 @@ identical to that of the \f3Elapsed\fP field.
NOTE: TotalCPU provides a measure of the task's parent process and
does not include CPU time of child processes.
-.TP
-\f3Tres\fP
-Trackable resources. These are the resources specified by the
-job at submission time. For more details see AccountingStorageTRES in slurm.conf.
-
.TP
\f3UID\fP
The user identifier of the user who ran the job.
diff --git a/src/sacct/print.c b/src/sacct/print.c
index a359d541378a475ec886a8edaae9b6916ea5fc9c..35b187958992541a4d5d901a602cf28552e92510 100644
--- a/src/sacct/print.c
+++ b/src/sacct/print.c
@@ -236,9 +236,10 @@ void print_fields(type_t type, void *object)
got_stats = true;
job_comp = NULL;
-
cpu_tres_rec_count = slurmdb_find_tres_count_in_string(
- job->tres_alloc_str, TRES_CPU);
+ (job->tres_alloc_str && job->tres_alloc_str[0]) ?
+ job->tres_alloc_str : job->tres_req_str,
+ TRES_CPU);
break;
case JOBSTEP:
job = step->job_ptr;
@@ -251,7 +252,10 @@ void print_fields(type_t type, void *object)
step->tres_alloc_str, TRES_CPU)))
step_cpu_tres_rec_count =
slurmdb_find_tres_count_in_string(
- job->tres_alloc_str, TRES_CPU);
+ (job->tres_alloc_str &&
+ job->tres_alloc_str[0]) ?
+ job->tres_alloc_str : job->tres_req_str,
+ TRES_CPU);
job_comp = NULL;
break;
@@ -314,6 +318,37 @@ void print_fields(type_t type, void *object)
tmp_char,
(curr_inx == field_count));
break;
+ case PRINT_ALLOC_NODES:
+ switch(type) {
+ case JOB:
+ tmp_int = job->alloc_nodes;
+ tmp_char = job->tres_alloc_str;
+ break;
+ case JOBSTEP:
+ tmp_int = step->nnodes;
+ tmp_char = step->tres_alloc_str;
+ break;
+ case JOBCOMP:
+ tmp_int = job_comp->node_cnt;
+ break;
+ default:
+ break;
+ }
+
+ if (!tmp_int && tmp_char) {
+ if ((tmp_uint64 =
+ slurmdb_find_tres_count_in_string(
+ tmp_char, TRES_NODE))
+ != INFINITE64)
+ tmp_int = tmp_uint64;
+ }
+ convert_num_unit((double)tmp_int, outbuf,
+ sizeof(outbuf), UNIT_NONE,
+ params.convert_flags);
+ field->print_routine(field,
+ outbuf,
+ (curr_inx == field_count));
+ break;
case PRINT_ACCOUNT:
switch(type) {
case JOB:
@@ -1412,24 +1447,28 @@ void print_fields(type_t type, void *object)
switch(type) {
case JOB:
tmp_int = job->alloc_nodes;
- tmp_char = job->nodes;
+ tmp_char = (job->tres_alloc_str &&
+ job->tres_alloc_str[0])
+ ? job->tres_alloc_str :
+ job->tres_req_str;
break;
case JOBSTEP:
tmp_int = step->nnodes;
- tmp_char = step->nodes;
+ tmp_char = step->tres_alloc_str;
break;
case JOBCOMP:
tmp_int = job_comp->node_cnt;
- tmp_char = job_comp->nodelist;
break;
default:
break;
}
- if (!tmp_int) {
- hostlist_t hl = hostlist_create(tmp_char);
- tmp_int = hostlist_count(hl);
- hostlist_destroy(hl);
+ if (!tmp_int && tmp_char) {
+ if ((tmp_uint64 =
+ slurmdb_find_tres_count_in_string(
+ tmp_char, TRES_NODE))
+ != INFINITE64)
+ tmp_int = tmp_uint64;
}
convert_num_unit((double)tmp_int, outbuf,
sizeof(outbuf), UNIT_NONE,
@@ -1667,6 +1706,38 @@ void print_fields(type_t type, void *object)
else
sprintf(outbuf+strlen(outbuf), "n");
}
+ field->print_routine(field,
+ outbuf,
+ (curr_inx == field_count));
+ break;
+ case PRINT_REQ_NODES:
+ switch(type) {
+ case JOB:
+ tmp_int = 0;
+ tmp_char = job->tres_req_str;
+ break;
+ case JOBSTEP:
+ tmp_int = step->nnodes;
+ tmp_char = step->tres_alloc_str;
+ break;
+ case JOBCOMP:
+ tmp_int = job_comp->node_cnt;
+ break;
+ default:
+ break;
+ }
+
+ if (!tmp_int && tmp_char) {
+ if ((tmp_uint64 =
+ slurmdb_find_tres_count_in_string(
+ tmp_char, TRES_NODE))
+ != INFINITE64)
+ tmp_int = tmp_uint64;
+ }
+ convert_num_unit((double)tmp_int, outbuf,
+ sizeof(outbuf), UNIT_NONE,
+ params.convert_flags);
+
field->print_routine(field,
outbuf,
(curr_inx == field_count));
diff --git a/src/sacct/sacct.c b/src/sacct/sacct.c
index c7a70b4a6ad745ef8e7cd08c5a836e5159332ca6..d00b2ba7ed61a80a71508f960a05157f70634c2c 100644
--- a/src/sacct/sacct.c
+++ b/src/sacct/sacct.c
@@ -46,9 +46,9 @@ sacct_parameters_t params;
print_field_t fields[] = {
{10, "AllocCPUS", print_fields_uint, PRINT_ALLOC_CPUS},
{12, "AllocGRES", print_fields_str, PRINT_ALLOC_GRES},
+ {10, "AllocNodes", print_fields_str, PRINT_ALLOC_NODES},
+ {10, "AllocTRES", print_fields_str, PRINT_TRESA},
{10, "Account", print_fields_str, PRINT_ACCOUNT},
- {10, "TRESAlloc", print_fields_str, PRINT_TRESA},
- {10, "TRESReq", print_fields_str, PRINT_TRESR},
{7, "AssocID", print_fields_uint, PRINT_ASSOCID},
{10, "AveCPU", print_fields_str, PRINT_AVECPU},
{10, "AveCPUFreq", print_fields_str, PRINT_ACT_CPUFREQ},
@@ -109,6 +109,8 @@ print_field_t fields[] = {
{8, "ReqCPUS", print_fields_uint, PRINT_REQ_CPUS},
{12, "ReqGRES", print_fields_str, PRINT_REQ_GRES},
{10, "ReqMem", print_fields_str, PRINT_REQ_MEM},
+ {8, "ReqNodes", print_fields_str, PRINT_REQ_NODES},
+ {10, "ReqTRES", print_fields_str, PRINT_TRESR},
{20, "Reservation", print_fields_str, PRINT_RESERVATION},
{8, "ReservationId", print_fields_uint, PRINT_RESERVATION_ID},
{10, "Reserved", print_fields_time_from_secs, PRINT_RESV},
diff --git a/src/sacct/sacct.h b/src/sacct/sacct.h
index 5bc804d5a18a6fd4274214b66dbdeb33fef7926a..3a56fedfbaf76a11f9682d53d0793fd3ccb65ba0 100644
--- a/src/sacct/sacct.h
+++ b/src/sacct/sacct.h
@@ -95,6 +95,7 @@ typedef enum {
PRINT_ACCOUNT,
PRINT_ALLOC_CPUS,
PRINT_ALLOC_GRES,
+ PRINT_ALLOC_NODES,
PRINT_TRESA,
PRINT_TRESR,
PRINT_ASSOCID,
@@ -154,6 +155,7 @@ typedef enum {
PRINT_REQ_CPUS,
PRINT_REQ_GRES,
PRINT_REQ_MEM,
+ PRINT_REQ_NODES,
PRINT_RESERVATION,
PRINT_RESERVATION_ID,
PRINT_RESV,
diff --git a/src/slurmctld/job_mgr.c b/src/slurmctld/job_mgr.c
index 093e0e888e59a25453fd75118c5cdf91c3dac114..bcaee1db9dff4c9f5112d2c87eadb7e60589d991 100644
--- a/src/slurmctld/job_mgr.c
+++ b/src/slurmctld/job_mgr.c
@@ -9686,14 +9686,18 @@ static int _update_job(struct job_record *job_ptr, job_desc_msg_t * job_specs,
resv_name = job_ptr->resv_name;
memset(&qos_rec, 0, sizeof(slurmdb_qos_rec_t));
- qos_rec.name = job_specs->qos;
+
+ /* If the qos is blank that means we want the default */
+ if (job_specs->qos[0])
+ qos_rec.name = job_specs->qos;
new_qos_ptr = _determine_and_validate_qos(
resv_name, job_ptr->assoc_ptr,
authorized, &qos_rec, &error_code, false);
if (error_code == SLURM_SUCCESS) {
info("%s: setting QOS to %s for job_id %u",
- __func__, job_specs->qos, job_ptr->job_id);
+ __func__, new_qos_ptr->name,
+ job_ptr->job_id);
if (job_ptr->qos_id != qos_rec.id) {
job_ptr->qos_id = qos_rec.id;
job_ptr->qos_ptr = new_qos_ptr;
diff --git a/src/slurmctld/step_mgr.c b/src/slurmctld/step_mgr.c
index 46aa9e317662230d03f4f7279f0cedc7a04375e9..b7f7f73049bfae4bf90a1ce4ea2537a4ad2859d3 100644
--- a/src/slurmctld/step_mgr.c
+++ b/src/slurmctld/step_mgr.c
@@ -2548,10 +2548,6 @@ step_create(job_step_create_request_msg_t *step_specs,
#endif
xfree(step_ptr->tres_alloc_str);
- xstrfmtcat(step_ptr->tres_alloc_str, "%s%u=%"PRIu64,
- step_ptr->tres_alloc_str ? "," : "",
- TRES_CPU, cpu_count);
-
tres_count = (uint64_t)step_ptr->pn_min_memory;
if (tres_count & MEM_PER_CPU) {
tres_count &= (~MEM_PER_CPU);
@@ -2559,9 +2555,12 @@ step_create(job_step_create_request_msg_t *step_specs,
} else
tres_count *= node_count;
- xstrfmtcat(step_ptr->tres_alloc_str, "%s%u=%"PRIu64,
+ xstrfmtcat(step_ptr->tres_alloc_str,
+ "%s%u=%"PRIu64",%u=%"PRIu64",%u=%u",
step_ptr->tres_alloc_str ? "," : "",
- TRES_MEM, tres_count);
+ TRES_CPU, cpu_count,
+ TRES_MEM, tres_count,
+ TRES_NODE, node_count);
if ((tmp_tres_str = gres_2_tres_str(step_ptr->gres_list, 0, true))) {
xstrfmtcat(step_ptr->tres_alloc_str, "%s%s",