From 0629201ade01009e82cea653dca342b840e4897c Mon Sep 17 00:00:00 2001
From: Danny Auble <da@schedmd.com>
Date: Wed, 12 Aug 2015 17:25:32 -0700
Subject: [PATCH] Added a bunch of new waiting codes dealing with TRES
---
slurm/slurm.h.in | 158 +++++++++++++--
src/common/slurm_protocol_defs.c | 195 +++++++++++++++++--
src/slurmctld/acct_policy.c | 318 ++++++++++++++++++++++++++++---
3 files changed, 602 insertions(+), 69 deletions(-)
diff --git a/slurm/slurm.h.in b/slurm/slurm.h.in
index eaa88700dcc..6befb6786e8 100644
--- a/slurm/slurm.h.in
+++ b/slurm/slurm.h.in
@@ -359,35 +359,36 @@ enum job_state_reason {
WAIT_DEP_INVALID, /* Dependency condition invalid or never
* satisfied
*/
- WAIT_QOS_GRP_CPU, /* QOS GrpCpus exceeded */
- WAIT_QOS_GRP_CPU_MIN, /* QOS GrpCPUMins exceeded */
- WAIT_QOS_GRP_CPU_RUN_MIN, /* QOS GrpCPURunMins exceeded */
+ WAIT_QOS_GRP_CPU, /* QOS GrpTRES exceeded (CPU) */
+ WAIT_QOS_GRP_CPU_MIN, /* QOS GrpTRESMins exceeded (CPU) */
+ WAIT_QOS_GRP_CPU_RUN_MIN, /* QOS GrpTRESRunMins exceeded (CPU) */
WAIT_QOS_GRP_JOB, /* QOS GrpJobs exceeded */
- WAIT_QOS_GRP_MEMORY, /* QOS GrpMemory exceeded */
- WAIT_QOS_GRP_NODES, /* QOS GrpNodes exceeded */
+ WAIT_QOS_GRP_MEM, /* QOS GrpTRES exceeded (Memory) */
+ WAIT_QOS_GRP_NODE, /* QOS GrpTRES exceeded (Node) */
WAIT_QOS_GRP_SUB_JOB, /* QOS GrpSubmitJobs exceeded */
WAIT_QOS_GRP_WALL, /* QOS GrpWall exceeded */
- WAIT_QOS_MAX_CPUS_PER_JOB, /* QOS MaxCpusPerJob exceeded */
- WAIT_QOS_MAX_CPU_MINS_PER_JOB,/* QOS MaxCpusMinsPerJob exceeded */
- WAIT_QOS_MAX_NODE_PER_JOB, /* QOS MaxNodesPerJob exceeded */
+ WAIT_QOS_MAX_CPU_PER_JOB, /* QOS MaxTRESPerJob exceeded (CPU) */
+ WAIT_QOS_MAX_CPU_MINS_PER_JOB,/* QOS MaxTRESMinsPerJob exceeded (CPU) */
+ WAIT_QOS_MAX_NODE_PER_JOB, /* QOS MaxTRESPerJob exceeded (Node) */
WAIT_QOS_MAX_WALL_PER_JOB, /* QOS MaxWallDurationPerJob exceeded */
- WAIT_QOS_MAX_CPU_PER_USER, /* QOS MaxCpusPerUser exceeded */
+ WAIT_QOS_MAX_CPU_PER_USER, /* QOS MaxTRESPerUser exceeded (CPU) */
WAIT_QOS_MAX_JOB_PER_USER, /* QOS MaxJobsPerUser exceeded */
- WAIT_QOS_MAX_NODE_PER_USER, /* QOS MaxNodesPerUser exceeded */
+ WAIT_QOS_MAX_NODE_PER_USER, /* QOS MaxTRESPerUser exceeded (Node) */
WAIT_QOS_MAX_SUB_JOB, /* QOS MaxSubmitJobsPerUser exceeded */
- WAIT_QOS_MIN_CPUS, /* QOS MinCPUsPerJob not reached */
- WAIT_ASSOC_GRP_CPU, /* ASSOC GrpCpus exceeded */
- WAIT_ASSOC_GRP_CPU_MIN, /* ASSOC GrpCPUMins exceeded */
- WAIT_ASSOC_GRP_CPU_RUN_MIN, /* ASSOC GrpCPURunMins exceeded */
+ WAIT_QOS_MIN_CPU, /* QOS MinTRESPerJob not reached (CPU) */
+ WAIT_ASSOC_GRP_CPU, /* ASSOC GrpTRES exceeded (CPU) */
+ WAIT_ASSOC_GRP_CPU_MIN, /* ASSOC GrpTRESMins exceeded (CPU) */
+ WAIT_ASSOC_GRP_CPU_RUN_MIN, /* ASSOC GrpTRESRunMins exceeded (CPU) */
WAIT_ASSOC_GRP_JOB, /* ASSOC GrpJobs exceeded */
- WAIT_ASSOC_GRP_MEMORY, /* ASSOC GrpMemory exceeded */
- WAIT_ASSOC_GRP_NODES, /* ASSOC GrpNodes exceeded */
+ WAIT_ASSOC_GRP_MEM, /* ASSOC GrpTRES exceeded (Memory) */
+ WAIT_ASSOC_GRP_NODE, /* ASSOC GrpTRES exceeded (Node) */
WAIT_ASSOC_GRP_SUB_JOB, /* ASSOC GrpSubmitJobs exceeded */
WAIT_ASSOC_GRP_WALL, /* ASSOC GrpWall exceeded */
WAIT_ASSOC_MAX_JOBS, /* ASSOC MaxJobs exceeded */
- WAIT_ASSOC_MAX_CPUS_PER_JOB, /* ASSOC MaxCpusPerJob exceeded */
- WAIT_ASSOC_MAX_CPU_MINS_PER_JOB,/* ASSOC MaxCpusMinsPerJob exceeded */
- WAIT_ASSOC_MAX_NODE_PER_JOB, /* ASSOC MaxNodesPerJob exceeded */
+ WAIT_ASSOC_MAX_CPU_PER_JOB, /* ASSOC MaxTRESPerJob exceeded (CPU) */
+ WAIT_ASSOC_MAX_CPU_MINS_PER_JOB,/* ASSOC MaxTRESMinsPerJob
+ * exceeded (CPU) */
+ WAIT_ASSOC_MAX_NODE_PER_JOB, /* ASSOC MaxTRESPerJob exceeded (NODE) */
WAIT_ASSOC_MAX_WALL_PER_JOB, /* ASSOC MaxWallDurationPerJob
* exceeded */
WAIT_ASSOC_MAX_SUB_JOB, /* ASSOC MaxSubmitJobsPerUser exceeded */
@@ -400,6 +401,125 @@ enum job_state_reason {
WAIT_POWER_NOT_AVAIL, /* not enough power available */
WAIT_POWER_RESERVED, /* job is waiting for available power
* because of power reservations */
+ WAIT_ASSOC_GRP_UNK, /* ASSOC GrpTRES exceeded
+ * (Unknown) */
+ WAIT_ASSOC_GRP_UNK_MIN, /* ASSOC GrpTRESMins exceeded
+ * (Unknown) */
+ WAIT_ASSOC_GRP_UNK_RUN_MIN, /* ASSOC GrpTRESRunMins exceeded
+ * (Unknown) */
+ WAIT_ASSOC_MAX_UNK_PER_JOB, /* ASSOC MaxTRESPerJob exceeded
+ * (Unknown) */
+ WAIT_ASSOC_MAX_UNK_MINS_PER_JOB,/* ASSOC MaxTRESMinsPerJob
+ * exceeded (Unknown) */
+ WAIT_ASSOC_GRP_MEM_MIN, /* ASSOC GrpTRESMins exceeded
+ * (Memory) */
+ WAIT_ASSOC_GRP_MEM_RUN_MIN, /* ASSOC GrpTRESRunMins exceeded
+ * (Memory) */
+ WAIT_ASSOC_MAX_MEM_PER_JOB, /* ASSOC MaxTRESPerJob exceeded (Memory) */
+ WAIT_ASSOC_MAX_MEM_MINS_PER_JOB,/* ASSOC MaxTRESMinsPerJob
+ * exceeded (Memory) */
+ WAIT_ASSOC_GRP_NODE_MIN, /* ASSOC GrpTRESMins exceeded (Node) */
+ WAIT_ASSOC_GRP_NODE_RUN_MIN, /* ASSOC GrpTRESRunMins exceeded (Node) */
+ WAIT_ASSOC_MAX_NODE_MINS_PER_JOB,/* ASSOC MaxTRESMinsPerJob
+ * exceeded (Node) */
+ WAIT_ASSOC_GRP_ENERGY, /* ASSOC GrpTRES exceeded
+ * (Energy) */
+ WAIT_ASSOC_GRP_ENERGY_MIN, /* ASSOC GrpTRESMins exceeded
+ * (Energy) */
+ WAIT_ASSOC_GRP_ENERGY_RUN_MIN, /* ASSOC GrpTRESRunMins exceeded
+ * (Energy) */
+ WAIT_ASSOC_MAX_ENERGY_PER_JOB, /* ASSOC MaxTRESPerJob exceeded
+ * (Energy) */
+ WAIT_ASSOC_MAX_ENERGY_MINS_PER_JOB,/* ASSOC MaxTRESMinsPerJob
+ * exceeded (Energy) */
+ WAIT_ASSOC_GRP_GRES, /* ASSOC GrpTRES exceeded (GRES) */
+ WAIT_ASSOC_GRP_GRES_MIN, /* ASSOC GrpTRESMins exceeded (GRES) */
+ WAIT_ASSOC_GRP_GRES_RUN_MIN, /* ASSOC GrpTRESRunMins exceeded (GRES) */
+ WAIT_ASSOC_MAX_GRES_PER_JOB, /* ASSOC MaxTRESPerJob exceeded (GRES) */
+ WAIT_ASSOC_MAX_GRES_MINS_PER_JOB,/* ASSOC MaxTRESMinsPerJob
+ * exceeded (GRES) */
+ WAIT_ASSOC_GRP_LIC, /* ASSOC GrpTRES exceeded
+ * (license) */
+ WAIT_ASSOC_GRP_LIC_MIN, /* ASSOC GrpTRESMins exceeded
+ * (license) */
+ WAIT_ASSOC_GRP_LIC_RUN_MIN, /* ASSOC GrpTRESRunMins exceeded
+ * (license) */
+ WAIT_ASSOC_MAX_LIC_PER_JOB, /* ASSOC MaxTRESPerJob exceeded
+ * (license) */
+ WAIT_ASSOC_MAX_LIC_MINS_PER_JOB,/* ASSOC MaxTRESMinsPerJob exceeded
+ * (license) */
+ WAIT_ASSOC_GRP_BB, /* ASSOC GrpTRES exceeded
+ * (burst buffer) */
+ WAIT_ASSOC_GRP_BB_MIN, /* ASSOC GrpTRESMins exceeded
+ * (burst buffer) */
+ WAIT_ASSOC_GRP_BB_RUN_MIN, /* ASSOC GrpTRESRunMins exceeded
+ * (burst buffer) */
+ WAIT_ASSOC_MAX_BB_PER_JOB, /* ASSOC MaxTRESPerJob exceeded
+ * (burst buffer) */
+ WAIT_ASSOC_MAX_BB_MINS_PER_JOB,/* ASSOC MaxTRESMinsPerJob exceeded
+ * (burst buffer) */
+ WAIT_QOS_GRP_UNK, /* QOS GrpTRES exceeded (Unknown) */
+ WAIT_QOS_GRP_UNK_MIN, /* QOS GrpTRESMins exceeded (Unknown) */
+ WAIT_QOS_GRP_UNK_RUN_MIN, /* QOS GrpTRESRunMins exceeded (Unknown) */
+ WAIT_QOS_MAX_UNK_PER_JOB, /* QOS MaxTRESPerJob exceeded (Unknown) */
+ WAIT_QOS_MAX_UNK_PER_USER, /* QOS MaxTRESPerUser exceeded (Unknown) */
+ WAIT_QOS_MAX_UNK_MINS_PER_JOB,/* QOS MaxTRESMinsPerJob
+ * exceeded (Unknown) */
+ WAIT_QOS_MIN_UNK, /* QOS MinTRESPerJob exceeded (Unknown) */
+ WAIT_QOS_GRP_MEM_MIN, /* QOS GrpTRESMins exceeded
+ * (Memory) */
+ WAIT_QOS_GRP_MEM_RUN_MIN, /* QOS GrpTRESRunMins exceeded
+ * (Memory) */
+ WAIT_QOS_MAX_MEM_MINS_PER_JOB,/* QOS MaxTRESMinsPerJob
+ * exceeded (Memory) */
+ WAIT_QOS_MAX_MEM_PER_JOB, /* QOS MaxTRESPerJob exceeded (CPU) */
+ WAIT_QOS_MAX_MEM_PER_USER, /* QOS MaxTRESPerUser exceeded (CPU) */
+ WAIT_QOS_MIN_MEM, /* QOS MinTRESPerJob not reached (Memory) */
+ WAIT_QOS_GRP_ENERGY, /* QOS GrpTRES exceeded (Energy) */
+ WAIT_QOS_GRP_ENERGY_MIN, /* QOS GrpTRESMins exceeded (Energy) */
+ WAIT_QOS_GRP_ENERGY_RUN_MIN, /* QOS GrpTRESRunMins exceeded (Energy) */
+ WAIT_QOS_MAX_ENERGY_PER_JOB, /* QOS MaxTRESPerJob exceeded (Energy) */
+ WAIT_QOS_MAX_ENERGY_PER_USER,/* QOS MaxTRESPerUser exceeded (Energy) */
+ WAIT_QOS_MAX_ENERGY_MINS_PER_JOB,/* QOS MaxTRESMinsPerJob
+ * exceeded (Energy) */
+ WAIT_QOS_MIN_ENERGY, /* QOS MinTRESPerJob not reached (Energy) */
+ WAIT_QOS_GRP_NODE_MIN, /* QOS GrpTRESMins exceeded (Node) */
+ WAIT_QOS_GRP_NODE_RUN_MIN, /* QOS GrpTRESRunMins exceeded (Node) */
+ WAIT_QOS_MAX_NODE_MINS_PER_JOB, /* QOS MaxTRESMinsPerJob
+ * exceeded (Node) */
+ WAIT_QOS_MIN_NODE, /* QOS MinTRESPerJob not reached (Node) */
+ WAIT_QOS_GRP_GRES, /* QOS GrpTRES exceeded (GRES) */
+ WAIT_QOS_GRP_GRES_MIN, /* QOS GrpTRESMins exceeded (GRES) */
+ WAIT_QOS_GRP_GRES_RUN_MIN, /* QOS GrpTRESRunMins exceeded (GRES) */
+ WAIT_QOS_MAX_GRES_PER_JOB, /* QOS MaxTRESPerJob exceeded (GRES) */
+ WAIT_QOS_MAX_GRES_PER_USER, /* QOS MaxTRESPerUser exceeded
+ * (GRES) */
+ WAIT_QOS_MAX_GRES_MINS_PER_JOB,/* QOS MaxTRESMinsPerJob
+ * exceeded (GRES) */
+ WAIT_QOS_MIN_GRES, /* QOS MinTRESPerJob not reached (CPU) */
+ WAIT_QOS_GRP_LIC, /* QOS GrpTRES exceeded (license) */
+ WAIT_QOS_GRP_LIC_MIN, /* QOS GrpTRESMins exceeded (license) */
+ WAIT_QOS_GRP_LIC_RUN_MIN, /* QOS GrpTRESRunMins exceeded (license) */
+ WAIT_QOS_MAX_LIC_PER_JOB, /* QOS MaxTRESPerJob exceeded (license) */
+ WAIT_QOS_MAX_LIC_PER_USER, /* QOS MaxTRESPerUser exceeded (license) */
+ WAIT_QOS_MAX_LIC_MINS_PER_JOB,/* QOS MaxTRESMinsPerJob exceeded
+ * (license) */
+ WAIT_QOS_MIN_LIC, /* QOS MinTRESPerJob not reached
+ * (license) */
+ WAIT_QOS_GRP_BB, /* QOS GrpTRES exceeded
+ * (burst buffer) */
+ WAIT_QOS_GRP_BB_MIN, /* QOS GrpTRESMins exceeded
+ * (burst buffer) */
+ WAIT_QOS_GRP_BB_RUN_MIN, /* QOS GrpTRESRunMins exceeded
+ * (burst buffer) */
+ WAIT_QOS_MAX_BB_PER_JOB, /* QOS MaxTRESPerJob exceeded
+ * (burst buffer) */
+ WAIT_QOS_MAX_BB_PER_USER, /* QOS MaxTRESPerUser exceeded
+ * (burst buffer) */
+ WAIT_QOS_MAX_BB_MINS_PER_JOB,/* QOS MaxTRESMinsPerJob exceeded
+ * (burst buffer) */
+ WAIT_QOS_MIN_BB, /* QOS MinTRESPerJob not reached
+ * (burst buffer) */
};
enum job_acct_types {
diff --git a/src/common/slurm_protocol_defs.c b/src/common/slurm_protocol_defs.c
index 420be2d2415..6567e0e9890 100644
--- a/src/common/slurm_protocol_defs.c
+++ b/src/common/slurm_protocol_defs.c
@@ -1335,32 +1335,32 @@ extern char *job_reason_string(enum job_state_reason inx)
return "QOSGrpCPURunMinsLimit";
case WAIT_QOS_GRP_JOB:
return"QOSGrpJobsLimit";
- case WAIT_QOS_GRP_MEMORY:
- return "QOSGrpMemoryLimit";
- case WAIT_QOS_GRP_NODES:
- return "QOSGrpNodesLimit";
+ case WAIT_QOS_GRP_MEM:
+ return "QOSGrpMemLimit";
+ case WAIT_QOS_GRP_NODE:
+ return "QOSGrpNodeLimit";
case WAIT_QOS_GRP_SUB_JOB:
return "QOSGrpSubmitJobsLimit";
case WAIT_QOS_GRP_WALL:
return "QOSGrpWallLimit";
- case WAIT_QOS_MAX_CPUS_PER_JOB:
- return "QOSMaxCpusPerJobLimit";
+ case WAIT_QOS_MAX_CPU_PER_JOB:
+ return "QOSMaxCpuPerJobLimit";
case WAIT_QOS_MAX_CPU_MINS_PER_JOB:
- return "QOSMaxCpusMinsPerJobLimit";
+ return "QOSMaxCpuMinsPerJobLimit";
case WAIT_QOS_MAX_NODE_PER_JOB:
- return "QOSMaxNodesPerJobLimit";
+ return "QOSMaxNodePerJobLimit";
case WAIT_QOS_MAX_WALL_PER_JOB:
return "QOSMaxWallDurationPerJobLimit";
case WAIT_QOS_MAX_CPU_PER_USER:
- return "QOSMaxCpusPerUserLimit";
+ return "QOSMaxCpuPerUserLimit";
case WAIT_QOS_MAX_JOB_PER_USER:
return "QOSMaxJobsPerUserLimit";
case WAIT_QOS_MAX_NODE_PER_USER:
- return "QOSMaxNodesPerUserLimit";
+ return "QOSMaxNodePerUserLimit";
case WAIT_QOS_MAX_SUB_JOB:
return "QOSMaxSubmitJobPerUserLimit";
- case WAIT_QOS_MIN_CPUS:
- return "QOSMinCPUsNotSatisfied";
+ case WAIT_QOS_MIN_CPU:
+ return "QOSMinCpuNotSatisfied";
case WAIT_ASSOC_GRP_CPU:
return "AssocGrpCpuLimit";
case WAIT_ASSOC_GRP_CPU_MIN:
@@ -1369,22 +1369,22 @@ extern char *job_reason_string(enum job_state_reason inx)
return "AssocGrpCPURunMinsLimit";
case WAIT_ASSOC_GRP_JOB:
return"AssocGrpJobsLimit";
- case WAIT_ASSOC_GRP_MEMORY:
- return "AssocGrpMemoryLimit";
- case WAIT_ASSOC_GRP_NODES:
- return "AssocGrpNodesLimit";
+ case WAIT_ASSOC_GRP_MEM:
+ return "AssocGrpMemLimit";
+ case WAIT_ASSOC_GRP_NODE:
+ return "AssocGrpNodeLimit";
case WAIT_ASSOC_GRP_SUB_JOB:
return "AssocGrpSubmitJobsLimit";
case WAIT_ASSOC_GRP_WALL:
return "AssocGrpWallLimit";
case WAIT_ASSOC_MAX_JOBS:
return "AssocMaxJobsLimit";
- case WAIT_ASSOC_MAX_CPUS_PER_JOB:
- return "AssocMaxCpusPerJobLimit";
+ case WAIT_ASSOC_MAX_CPU_PER_JOB:
+ return "AssocMaxCpuPerJobLimit";
case WAIT_ASSOC_MAX_CPU_MINS_PER_JOB:
- return "AssocMaxCpusMinsPerJobLimit";
+ return "AssocMaxCpuMinsPerJobLimit";
case WAIT_ASSOC_MAX_NODE_PER_JOB:
- return "AssocMaxNodesPerJobLimit";
+ return "AssocMaxNodePerJobLimit";
case WAIT_ASSOC_MAX_WALL_PER_JOB:
return "AssocMaxWallDurationPerJobLimit";
case WAIT_ASSOC_MAX_SUB_JOB:
@@ -1403,6 +1403,161 @@ extern char *job_reason_string(enum job_state_reason inx)
return "PowerNotAvail";
case WAIT_POWER_RESERVED:
return "PowerReserved";
+ case WAIT_ASSOC_GRP_UNK:
+ return "AssocGrpUnknown";
+ case WAIT_ASSOC_GRP_UNK_MIN:
+ return "AssocGrpUnknownMin";
+ case WAIT_ASSOC_GRP_UNK_RUN_MIN:
+ return "AssocGrpUnknownRunMin";
+ case WAIT_ASSOC_MAX_UNK_PER_JOB:
+ return "AssocMaxUnknownPerJob";
+ case WAIT_ASSOC_MAX_UNK_MINS_PER_JOB:
+ return "AssocMaxUnknownMinPerJob";
+ case WAIT_ASSOC_GRP_MEM_MIN:
+ return "AssocGrpMemMin";
+ case WAIT_ASSOC_GRP_MEM_RUN_MIN:
+ return "AssocGrpMemRunMin";
+ case WAIT_ASSOC_MAX_MEM_PER_JOB:
+ return "AssocMaxMemPerJob";
+ case WAIT_ASSOC_MAX_MEM_MINS_PER_JOB:
+ return "AssocMaxMemMinPerJob";
+ case WAIT_ASSOC_GRP_NODE_MIN:
+ return "AssocGrpNodeMin";
+ case WAIT_ASSOC_GRP_NODE_RUN_MIN:
+ return "AssocGrpNodeRunMin";
+ case WAIT_ASSOC_MAX_NODE_MINS_PER_JOB:
+ return "AssocMaxNodeMinPerJob";
+ case WAIT_ASSOC_GRP_ENERGY:
+ return "AssocGrpEnergy";
+ case WAIT_ASSOC_GRP_ENERGY_MIN:
+ return "AssocGrpEnergyMin";
+ case WAIT_ASSOC_GRP_ENERGY_RUN_MIN:
+ return "AssocGrpEnergyRunMin";
+ case WAIT_ASSOC_MAX_ENERGY_PER_JOB:
+ return "AssocMaxEnergyPerJob";
+ case WAIT_ASSOC_MAX_ENERGY_MINS_PER_JOB:
+ return "AssocMaxEnergyMinPerJob";
+ case WAIT_ASSOC_GRP_GRES:
+ return "AssocGrpGRES";
+ case WAIT_ASSOC_GRP_GRES_MIN:
+ return "AssocGrpGRESMin";
+ case WAIT_ASSOC_GRP_GRES_RUN_MIN:
+ return "AssocGrpGRESRunMin";
+ case WAIT_ASSOC_MAX_GRES_PER_JOB:
+ return "AssocMaxGRESPerJob";
+ case WAIT_ASSOC_MAX_GRES_MINS_PER_JOB:
+ return "AssocMaxGRESMinPerJob";
+ case WAIT_ASSOC_GRP_LIC:
+ return "AssocGrpLicense";
+ case WAIT_ASSOC_GRP_LIC_MIN:
+ return "AssocGrpLicenseMin";
+ case WAIT_ASSOC_GRP_LIC_RUN_MIN:
+ return "AssocGrpLicenseRunMin";
+ case WAIT_ASSOC_MAX_LIC_PER_JOB:
+ return "AssocMaxLicensePerJob";
+ case WAIT_ASSOC_MAX_LIC_MINS_PER_JOB:
+ return "AssocMaxLicenseMinPerJob";
+ case WAIT_ASSOC_GRP_BB:
+ return "AssocGrpBB";
+ case WAIT_ASSOC_GRP_BB_MIN:
+ return "AssocGrpBBMin";
+ case WAIT_ASSOC_GRP_BB_RUN_MIN:
+ return "AssocGrpBBRunMin";
+ case WAIT_ASSOC_MAX_BB_PER_JOB:
+ return "AssocMaxBBPerJob";
+ case WAIT_ASSOC_MAX_BB_MINS_PER_JOB:
+ return "AssocMaxBBMinPerJob";
+
+ case WAIT_QOS_GRP_UNK:
+ return "QOSGrpUnknown";
+ case WAIT_QOS_GRP_UNK_MIN:
+ return "QOSGrpUnknownMin";
+ case WAIT_QOS_GRP_UNK_RUN_MIN:
+ return "QOSGrpUnknownRunMin";
+ case WAIT_QOS_MAX_UNK_PER_JOB:
+ return "QOSMaxUnknownPerJob";
+ case WAIT_QOS_MAX_UNK_PER_USER:
+ return "QOSMaxUnknownPerUser";
+ case WAIT_QOS_MAX_UNK_MINS_PER_JOB:
+ return "QOSMaxUnknownMinPerJob";
+ case WAIT_QOS_MIN_UNK:
+ return "QOSMinUnknown";
+ case WAIT_QOS_GRP_MEM_MIN:
+ return "QOSGrpMemoryMin";
+ case WAIT_QOS_GRP_MEM_RUN_MIN:
+ return "QOSGrpMemoryRunMin";
+ case WAIT_QOS_MAX_MEM_PER_JOB:
+ return "QOSMaxMemoryPerJob";
+ case WAIT_QOS_MAX_MEM_PER_USER:
+ return "QOSMaxMemoryPerUser";
+ case WAIT_QOS_MAX_MEM_MINS_PER_JOB:
+ return "QOSMaxMemoryMinPerJob";
+ case WAIT_QOS_MIN_MEM:
+ return "QOSMinMemory";
+ case WAIT_QOS_GRP_NODE_MIN:
+ return "QOSGrpNodeMin";
+ case WAIT_QOS_GRP_NODE_RUN_MIN:
+ return "QOSGrpNodeRunMin";
+ case WAIT_QOS_MAX_NODE_MINS_PER_JOB:
+ return "QOSMaxNodeMinPerJob";
+ case WAIT_QOS_MIN_NODE:
+ return "QOSMinNode";
+ case WAIT_QOS_GRP_ENERGY:
+ return "QOSGrpEnergy";
+ case WAIT_QOS_GRP_ENERGY_MIN:
+ return "QOSGrpEnergyMin";
+ case WAIT_QOS_GRP_ENERGY_RUN_MIN:
+ return "QOSGrpEnergyRunMin";
+ case WAIT_QOS_MAX_ENERGY_PER_JOB:
+ return "QOSMaxEnergyPerJob";
+ case WAIT_QOS_MAX_ENERGY_PER_USER:
+ return "QOSMaxEnergyPerUser";
+ case WAIT_QOS_MAX_ENERGY_MINS_PER_JOB:
+ return "QOSMaxEnergyMinPerJob";
+ case WAIT_QOS_MIN_ENERGY:
+ return "QOSMinEnergy";
+ case WAIT_QOS_GRP_GRES:
+ return "QOSGrpGRES";
+ case WAIT_QOS_GRP_GRES_MIN:
+ return "QOSGrpGRESMin";
+ case WAIT_QOS_GRP_GRES_RUN_MIN:
+ return "QOSGrpGRESRunMin";
+ case WAIT_QOS_MAX_GRES_PER_JOB:
+ return "QOSMaxGRESPerJob";
+ case WAIT_QOS_MAX_GRES_PER_USER:
+ return "QOSMaxGRESPerUser";
+ case WAIT_QOS_MAX_GRES_MINS_PER_JOB:
+ return "QOSMaxGRESMinPerJob";
+ case WAIT_QOS_MIN_GRES:
+ return "QOSMinGRES";
+ case WAIT_QOS_GRP_LIC:
+ return "QOSGrpLicense";
+ case WAIT_QOS_GRP_LIC_MIN:
+ return "QOSGrpLicenseMin";
+ case WAIT_QOS_GRP_LIC_RUN_MIN:
+ return "QOSGrpLicenseRunMin";
+ case WAIT_QOS_MAX_LIC_PER_JOB:
+ return "QOSMaxLicensePerJob";
+ case WAIT_QOS_MAX_LIC_PER_USER:
+ return "QOSMaxLicensePerUser";
+ case WAIT_QOS_MAX_LIC_MINS_PER_JOB:
+ return "QOSMaxLicenseMinPerJob";
+ case WAIT_QOS_MIN_LIC:
+ return "QOSMinLicense";
+ case WAIT_QOS_GRP_BB:
+ return "QOSGrpBB";
+ case WAIT_QOS_GRP_BB_MIN:
+ return "QOSGrpBBMin";
+ case WAIT_QOS_GRP_BB_RUN_MIN:
+ return "QOSGrpBBRunMin";
+ case WAIT_QOS_MAX_BB_PER_JOB:
+ return "QOSMaxBBPerJob";
+ case WAIT_QOS_MAX_BB_PER_USER:
+ return "QOSMaxBBPerUser";
+ case WAIT_QOS_MAX_BB_MINS_PER_JOB:
+ return "AssocMaxBBMinPerJob";
+ case WAIT_QOS_MIN_BB:
+ return "QOSMinBB";
default:
snprintf(val, sizeof(val), "%d", inx);
return val;
diff --git a/src/slurmctld/acct_policy.c b/src/slurmctld/acct_policy.c
index 2ae0e075d5a..1eee69a091c 100644
--- a/src/slurmctld/acct_policy.c
+++ b/src/slurmctld/acct_policy.c
@@ -59,6 +59,231 @@ enum {
ACCT_POLICY_JOB_FINI
};
+static int get_tres_state_reason(int tres_pos, int unk_reason)
+{
+ switch (tres_pos) {
+ case TRES_ARRAY_CPU:
+ switch (unk_reason) {
+ case WAIT_ASSOC_GRP_UNK:
+ return WAIT_ASSOC_GRP_CPU;
+ case WAIT_ASSOC_GRP_UNK_MIN:
+ return WAIT_ASSOC_GRP_CPU_MIN;
+ case WAIT_ASSOC_GRP_UNK_RUN_MIN:
+ return WAIT_ASSOC_GRP_CPU_RUN_MIN;
+ case WAIT_ASSOC_MAX_UNK_PER_JOB:
+ return WAIT_ASSOC_MAX_CPU_PER_JOB;
+ case WAIT_ASSOC_MAX_UNK_MINS_PER_JOB:
+ return WAIT_ASSOC_MAX_CPU_MINS_PER_JOB;
+ case WAIT_QOS_GRP_UNK:
+ return WAIT_QOS_GRP_CPU;
+ case WAIT_QOS_GRP_UNK_MIN:
+ return WAIT_QOS_GRP_CPU_MIN;
+ case WAIT_QOS_GRP_UNK_RUN_MIN:
+ return WAIT_QOS_GRP_CPU_RUN_MIN;
+ case WAIT_QOS_MAX_UNK_PER_JOB:
+ return WAIT_QOS_MAX_CPU_PER_JOB;
+ case WAIT_QOS_MAX_UNK_PER_USER:
+ return WAIT_QOS_MAX_CPU_PER_USER;
+ case WAIT_QOS_MAX_UNK_MINS_PER_JOB:
+ return WAIT_QOS_MAX_CPU_MINS_PER_JOB;
+ case WAIT_QOS_MIN_UNK:
+ return WAIT_QOS_MIN_CPU;
+ default:
+ return unk_reason;
+ break;
+ }
+ break;
+ case TRES_ARRAY_MEM:
+ switch (unk_reason) {
+ case WAIT_ASSOC_GRP_UNK:
+ return WAIT_ASSOC_GRP_MEM;
+ case WAIT_ASSOC_GRP_UNK_MIN:
+ return WAIT_ASSOC_GRP_MEM_MIN;
+ case WAIT_ASSOC_GRP_UNK_RUN_MIN:
+ return WAIT_ASSOC_GRP_MEM_RUN_MIN;
+ case WAIT_ASSOC_MAX_UNK_PER_JOB:
+ return WAIT_ASSOC_MAX_MEM_PER_JOB;
+ case WAIT_ASSOC_MAX_UNK_MINS_PER_JOB:
+ return WAIT_ASSOC_MAX_MEM_MINS_PER_JOB;
+ case WAIT_QOS_GRP_UNK:
+ return WAIT_QOS_GRP_MEM;
+ case WAIT_QOS_GRP_UNK_MIN:
+ return WAIT_QOS_GRP_MEM_MIN;
+ case WAIT_QOS_GRP_UNK_RUN_MIN:
+ return WAIT_QOS_GRP_MEM_RUN_MIN;
+ case WAIT_QOS_MAX_UNK_PER_JOB:
+ return WAIT_QOS_MAX_MEM_PER_JOB;
+ case WAIT_QOS_MAX_UNK_PER_USER:
+ return WAIT_QOS_MAX_MEM_PER_USER;
+ case WAIT_QOS_MAX_UNK_MINS_PER_JOB:
+ return WAIT_QOS_MAX_MEM_MINS_PER_JOB;
+ case WAIT_QOS_MIN_UNK:
+ return WAIT_QOS_MIN_MEM;
+ default:
+ return unk_reason;
+ break;
+ }
+ break;
+ case TRES_ARRAY_ENEGRY:
+ switch (unk_reason) {
+ case WAIT_ASSOC_GRP_UNK:
+ return WAIT_ASSOC_GRP_ENERGY;
+ case WAIT_ASSOC_GRP_UNK_MIN:
+ return WAIT_ASSOC_GRP_ENERGY_MIN;
+ case WAIT_ASSOC_GRP_UNK_RUN_MIN:
+ return WAIT_ASSOC_GRP_ENERGY_RUN_MIN;
+ case WAIT_ASSOC_MAX_UNK_PER_JOB:
+ return WAIT_ASSOC_MAX_ENERGY_PER_JOB;
+ case WAIT_ASSOC_MAX_UNK_MINS_PER_JOB:
+ return WAIT_ASSOC_MAX_ENERGY_MINS_PER_JOB;
+ case WAIT_QOS_GRP_UNK:
+ return WAIT_QOS_GRP_ENERGY;
+ case WAIT_QOS_GRP_UNK_MIN:
+ return WAIT_QOS_GRP_ENERGY_MIN;
+ case WAIT_QOS_GRP_UNK_RUN_MIN:
+ return WAIT_QOS_GRP_ENERGY_RUN_MIN;
+ case WAIT_QOS_MAX_UNK_PER_JOB:
+ return WAIT_QOS_MAX_ENERGY_PER_JOB;
+ case WAIT_QOS_MAX_UNK_PER_USER:
+ return WAIT_QOS_MAX_ENERGY_PER_USER;
+ case WAIT_QOS_MAX_UNK_MINS_PER_JOB:
+ return WAIT_QOS_MAX_ENERGY_MINS_PER_JOB;
+ case WAIT_QOS_MIN_UNK:
+ return WAIT_QOS_MIN_ENERGY;
+ default:
+ return unk_reason;
+ break;
+ }
+ break;
+ case TRES_ARRAY_NODE:
+ switch (unk_reason) {
+ case WAIT_ASSOC_GRP_UNK:
+ return WAIT_ASSOC_GRP_NODE;
+ case WAIT_ASSOC_GRP_UNK_MIN:
+ return WAIT_ASSOC_GRP_NODE_MIN;
+ case WAIT_ASSOC_GRP_UNK_RUN_MIN:
+ return WAIT_ASSOC_GRP_NODE_RUN_MIN;
+ case WAIT_ASSOC_MAX_UNK_PER_JOB:
+ return WAIT_ASSOC_MAX_NODE_PER_JOB;
+ case WAIT_ASSOC_MAX_UNK_MINS_PER_JOB:
+ return WAIT_ASSOC_MAX_NODE_MINS_PER_JOB;
+ case WAIT_QOS_GRP_UNK:
+ return WAIT_QOS_GRP_NODE;
+ case WAIT_QOS_GRP_UNK_MIN:
+ return WAIT_QOS_GRP_NODE_MIN;
+ case WAIT_QOS_GRP_UNK_RUN_MIN:
+ return WAIT_QOS_GRP_NODE_RUN_MIN;
+ case WAIT_QOS_MAX_UNK_PER_JOB:
+ return WAIT_QOS_MAX_NODE_PER_JOB;
+ case WAIT_QOS_MAX_UNK_PER_USER:
+ return WAIT_QOS_MAX_NODE_PER_USER;
+ case WAIT_QOS_MAX_UNK_MINS_PER_JOB:
+ return WAIT_QOS_MAX_NODE_MINS_PER_JOB;
+ case WAIT_QOS_MIN_UNK:
+ return WAIT_QOS_MIN_NODE;
+ default:
+ return unk_reason;
+ break;
+ }
+ break;
+ default:
+ if (!xstrcmp("gres", assoc_mgr_tres_array[tres_pos]->type))
+ switch (unk_reason) {
+ case WAIT_ASSOC_GRP_UNK:
+ return WAIT_ASSOC_GRP_GRES;
+ case WAIT_ASSOC_GRP_UNK_MIN:
+ return WAIT_ASSOC_GRP_GRES_MIN;
+ case WAIT_ASSOC_GRP_UNK_RUN_MIN:
+ return WAIT_ASSOC_GRP_GRES_RUN_MIN;
+ case WAIT_ASSOC_MAX_UNK_PER_JOB:
+ return WAIT_ASSOC_MAX_GRES_PER_JOB;
+ case WAIT_ASSOC_MAX_UNK_MINS_PER_JOB:
+ return WAIT_ASSOC_MAX_GRES_MINS_PER_JOB;
+ case WAIT_QOS_GRP_UNK:
+ return WAIT_QOS_GRP_GRES;
+ case WAIT_QOS_GRP_UNK_MIN:
+ return WAIT_QOS_GRP_GRES_MIN;
+ case WAIT_QOS_GRP_UNK_RUN_MIN:
+ return WAIT_QOS_GRP_GRES_RUN_MIN;
+ case WAIT_QOS_MAX_UNK_PER_JOB:
+ return WAIT_QOS_MAX_GRES_PER_JOB;
+ case WAIT_QOS_MAX_UNK_PER_USER:
+ return WAIT_QOS_MAX_GRES_PER_USER;
+ case WAIT_QOS_MAX_UNK_MINS_PER_JOB:
+ return WAIT_QOS_MAX_GRES_MINS_PER_JOB;
+ case WAIT_QOS_MIN_UNK:
+ return WAIT_QOS_MIN_GRES;
+ default:
+ return unk_reason;
+ break;
+ }
+ else if (!xstrcmp("license",
+ assoc_mgr_tres_array[tres_pos]->type))
+ switch (unk_reason) {
+ case WAIT_ASSOC_GRP_UNK:
+ return WAIT_ASSOC_GRP_LIC;
+ case WAIT_ASSOC_GRP_UNK_MIN:
+ return WAIT_ASSOC_GRP_LIC_MIN;
+ case WAIT_ASSOC_GRP_UNK_RUN_MIN:
+ return WAIT_ASSOC_GRP_LIC_RUN_MIN;
+ case WAIT_ASSOC_MAX_UNK_PER_JOB:
+ return WAIT_ASSOC_MAX_LIC_PER_JOB;
+ case WAIT_ASSOC_MAX_UNK_MINS_PER_JOB:
+ return WAIT_ASSOC_MAX_LIC_MINS_PER_JOB;
+ case WAIT_QOS_GRP_UNK:
+ return WAIT_QOS_GRP_LIC;
+ case WAIT_QOS_GRP_UNK_MIN:
+ return WAIT_QOS_GRP_LIC_MIN;
+ case WAIT_QOS_GRP_UNK_RUN_MIN:
+ return WAIT_QOS_GRP_LIC_RUN_MIN;
+ case WAIT_QOS_MAX_UNK_PER_JOB:
+ return WAIT_QOS_MAX_LIC_PER_JOB;
+ case WAIT_QOS_MAX_UNK_PER_USER:
+ return WAIT_QOS_MAX_LIC_PER_USER;
+ case WAIT_QOS_MAX_UNK_MINS_PER_JOB:
+ return WAIT_QOS_MAX_LIC_MINS_PER_JOB;
+ case WAIT_QOS_MIN_UNK:
+ return WAIT_QOS_MIN_LIC;
+ default:
+ return unk_reason;
+ break;
+ }
+ else if (!xstrcmp("bb", assoc_mgr_tres_array[tres_pos]->type))
+ switch (unk_reason) {
+ case WAIT_ASSOC_GRP_UNK:
+ return WAIT_ASSOC_GRP_BB;
+ case WAIT_ASSOC_GRP_UNK_MIN:
+ return WAIT_ASSOC_GRP_BB_MIN;
+ case WAIT_ASSOC_GRP_UNK_RUN_MIN:
+ return WAIT_ASSOC_GRP_BB_RUN_MIN;
+ case WAIT_ASSOC_MAX_UNK_PER_JOB:
+ return WAIT_ASSOC_MAX_BB_PER_JOB;
+ case WAIT_ASSOC_MAX_UNK_MINS_PER_JOB:
+ return WAIT_ASSOC_MAX_BB_MINS_PER_JOB;
+ case WAIT_QOS_GRP_UNK:
+ return WAIT_QOS_GRP_BB;
+ case WAIT_QOS_GRP_UNK_MIN:
+ return WAIT_QOS_GRP_BB_MIN;
+ case WAIT_QOS_GRP_UNK_RUN_MIN:
+ return WAIT_QOS_GRP_BB_RUN_MIN;
+ case WAIT_QOS_MAX_UNK_PER_JOB:
+ return WAIT_QOS_MAX_BB_PER_JOB;
+ case WAIT_QOS_MAX_UNK_PER_USER:
+ return WAIT_QOS_MAX_BB_PER_USER;
+ case WAIT_QOS_MAX_UNK_MINS_PER_JOB:
+ return WAIT_QOS_MAX_BB_MINS_PER_JOB;
+ case WAIT_QOS_MIN_UNK:
+ return WAIT_QOS_MIN_BB;
+ default:
+ return unk_reason;
+ break;
+ }
+ break;
+ }
+
+ return unk_reason;
+}
+
static void _set_qos_order(struct job_record *job_ptr,
slurmdb_qos_rec_t **qos_ptr_1,
slurmdb_qos_rec_t **qos_ptr_2)
@@ -745,7 +970,8 @@ static int _qos_policy_validate(job_desc_msg_t *job_desc,
if (job_desc->tres_req_cnt[tres_pos] >
qos_ptr->max_tres_pu_ctld[tres_pos]) {
if (reason)
- *reason = WAIT_QOS_MAX_CPU_PER_USER;
+ *reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_MAX_UNK_PER_USER);
debug2("job submit for user %s(%u): "
"min tres(%s%s%s) request %"PRIu64" exceeds "
@@ -764,7 +990,8 @@ static int _qos_policy_validate(job_desc_msg_t *job_desc,
} else if (job_desc->tres_req_cnt[tres_pos] >
qos_ptr->grp_tres_ctld[tres_pos]) {
if (reason)
- *reason = WAIT_QOS_GRP_CPU;
+ *reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_GRP_UNK);
debug2("job submit for user %s(%u): "
"min tres(%s%s%s) request %"PRIu64" exceeds "
@@ -825,6 +1052,9 @@ static int _qos_policy_validate(job_desc_msg_t *job_desc,
qos_out_ptr->max_tres_mins_pj_ctld,
&acct_policy_limit_set->time,
strict_checking)) {
+ if (reason)
+ *reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_MAX_UNK_PER_JOB);
debug2("job submit for user %s(%u): "
"tres(%s%s%s) time limit request %"PRIu64" "
"exceeds max per-job limit %"PRIu64" "
@@ -839,7 +1069,8 @@ static int _qos_policy_validate(job_desc_msg_t *job_desc,
job_desc->tres_req_cnt[tres_pos]),
qos_ptr->max_tres_mins_pj_ctld[tres_pos],
qos_ptr->name);
-
+ rc = false;
+ goto end_it;
}
if ((qos_out_ptr->max_wall_pj == INFINITE) &&
@@ -897,7 +1128,8 @@ static int _qos_policy_validate(job_desc_msg_t *job_desc,
acct_policy_limit_set->tres,
strict_checking, 1)) {
if (reason)
- *reason = WAIT_QOS_MAX_CPUS_PER_JOB;
+ *reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_MAX_UNK_PER_JOB);
debug2("job submit for user %s(%u): "
"min tres(%s%s%s) request %"PRIu64" exceeds "
@@ -957,7 +1189,8 @@ static int _qos_policy_validate(job_desc_msg_t *job_desc,
acct_policy_limit_set->tres,
strict_checking, 0)) {
if (reason)
- *reason = WAIT_QOS_MIN_CPUS;
+ *reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_MIN_UNK);
debug2("job submit for user %s(%u): "
"min tres(%s%s%s) request %"PRIu64" exceeds "
@@ -1170,7 +1403,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
switch (i) {
case 1:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_GRP_CPU_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_GRP_UNK_MIN);
debug2("Job %u being held, "
"QOS %s group max tres(%s%s%s) minutes limit "
"of %"PRIu64" is already at or exceeded with %"PRIu64,
@@ -1187,7 +1421,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
break;
case 2:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_GRP_CPU_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_GRP_UNK_MIN);
debug2("Job %u being held, "
"the job is requesting more than allowed with QOS %s's "
"group max tres(%s%s%s) minutes of %"PRIu64" "
@@ -1212,7 +1447,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
* being killed
*/
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_GRP_CPU_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_GRP_UNK_MIN);
debug2("Job %u being held, "
"the job is at or exceeds QOS %s's "
"group max tres(%s%s%s) minutes of %"PRIu64" "
@@ -1255,7 +1491,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
break;
case 2:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_GRP_CPU;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_GRP_UNK);
debug2("job %u is being held, "
"QOS %s min tres(%s%s%s) request %"PRIu64" exceeds "
"group max tres limit %"PRIu64,
@@ -1272,7 +1509,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
break;
case 3:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_GRP_CPU;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_GRP_UNK);
debug2("job %u being held, "
"if allowed the job request will exceed "
"QOS %s group max tres(%s%s%s) limit "
@@ -1307,7 +1545,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
break;
case 2:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_GRP_CPU_RUN_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_GRP_UNK_RUN_MIN);
debug2("job %u is being held, "
"QOS %s group max running tres(%s%s%s) minutes "
"limit %"PRIu64" is already full with %"PRIu64,
@@ -1324,7 +1563,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
break;
case 3:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_GRP_CPU_RUN_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_GRP_UNK_RUN_MIN);
debug2("job %u being held, "
"if allowed the job request will exceed "
"QOS %s group max running tres(%s%s%s) minutes "
@@ -1359,7 +1599,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
job_ptr->limit_set.tres,
1, 1)) {
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_MAX_CPU_MINS_PER_JOB;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_MAX_UNK_MINS_PER_JOB);
debug2("Job %u being held, "
"the job is requesting more than allowed with QOS %s's "
"max tres(%s%s%s) minutes of %"PRIu64" "
@@ -1385,7 +1626,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
job_ptr->limit_set.tres,
1, 1)) {
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_MAX_CPUS_PER_JOB;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_MAX_UNK_PER_JOB);
debug2("job %u is being held, "
"QOS %s min tres(%s%s%s) per job "
"request %"PRIu64" exceeds "
@@ -1411,7 +1653,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
job_ptr->limit_set.tres,
1, 0)) {
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_MIN_CPUS;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_MIN_UNK);
debug2("job %u is being held, "
"QOS %s min tres(%s%s%s) per job "
"request %"PRIu64" exceeds "
@@ -1442,7 +1685,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
* TRES limit for the given QOS
*/
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_MAX_CPU_PER_USER;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_MAX_UNK_PER_USER);
debug2("job %u is being held, "
"QOS %s min tres(%s%s%s) "
"request %"PRIu64" exceeds "
@@ -1463,7 +1707,8 @@ static int _qos_job_runnable_post_select(struct job_record *job_ptr,
* the QOS per-user TRES limit with their
* current usage */
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_QOS_MAX_CPU_PER_USER;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_QOS_MAX_UNK_PER_USER);
debug2("job %u being held, "
"if allowed the job request will exceed "
"QOS %s max tres(%s%s%s) per user limit "
@@ -1792,7 +2037,9 @@ extern bool acct_policy_validate(job_desc_msg_t *job_desc,
we want to send back.
*/
if (reason)
- *reason = WAIT_ASSOC_GRP_CPU;
+ *reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_GRP_UNK);
+
debug2("job submit for user %s(%u): "
"min tres(%s%s%s) request %"PRIu64" exceeds "
"group max tres limit %"PRIu64" for account %s",
@@ -1859,7 +2106,9 @@ extern bool acct_policy_validate(job_desc_msg_t *job_desc,
we want to send back.
*/
if (reason)
- *reason = WAIT_ASSOC_MAX_CPUS_PER_JOB;
+ *reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_MAX_UNK_PER_JOB);
+
debug2("job submit for user %s(%u): "
"min tres(%s%s%s) request %"PRIu64" exceeds "
"max tres limit %"PRIu64" for account %s",
@@ -2244,7 +2493,8 @@ extern bool acct_policy_job_runnable_post_select(
switch (i) {
case 1:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_ASSOC_GRP_CPU_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_GRP_UNK_MIN);
debug2("Job %u being held, "
"assoc %u(%s/%s/%s) group max tres(%s%s%s) "
"minutes limit of %"PRIu64" is already at or "
@@ -2263,7 +2513,8 @@ extern bool acct_policy_job_runnable_post_select(
break;
case 2:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_ASSOC_GRP_CPU_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_GRP_UNK_MIN);
debug2("Job %u being held, "
"the job is requesting more than allowed "
"with assoc %u(%s/%s/%s) "
@@ -2290,7 +2541,8 @@ extern bool acct_policy_job_runnable_post_select(
* being killed
*/
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_ASSOC_GRP_CPU_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_GRP_UNK_MIN);
debug2("Job %u being held, "
"the job is at or exceeds assoc %u(%s/%s/%s) "
"group max tres(%s%s%s) minutes of %"PRIu64" "
@@ -2332,7 +2584,8 @@ extern bool acct_policy_job_runnable_post_select(
break;
case 2:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_ASSOC_GRP_CPU;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_GRP_UNK);
debug2("job %u is being held, "
"assoc %u(%s/%s/%s) min tres(%s%s%s) "
"request %"PRIu64" exceeds "
@@ -2351,7 +2604,8 @@ extern bool acct_policy_job_runnable_post_select(
break;
case 3:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_ASSOC_GRP_CPU;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_GRP_UNK);
debug2("job %u being held, "
"if allowed the job request will exceed "
"assoc %u(%s/%s/%s) group max "
@@ -2388,7 +2642,8 @@ extern bool acct_policy_job_runnable_post_select(
break;
case 2:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_ASSOC_GRP_CPU_RUN_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_GRP_UNK_RUN_MIN);
debug2("job %u is being held, "
"assoc %u(%s/%s/%s) group max running "
"tres(%s%s%s) minutes limit %"PRIu64
@@ -2407,7 +2662,8 @@ extern bool acct_policy_job_runnable_post_select(
break;
case 3:
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_ASSOC_GRP_CPU_RUN_MIN;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_GRP_UNK_RUN_MIN);
debug2("job %u being held, "
"if allowed the job request will exceed "
"assoc %u(%s/%s/%s) group max running "
@@ -2453,7 +2709,8 @@ extern bool acct_policy_job_runnable_post_select(
job_ptr->limit_set.tres,
1, 0, 1)) {
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_ASSOC_MAX_CPU_MINS_PER_JOB;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_MAX_UNK_MINS_PER_JOB);
debug2("Job %u being held, "
"the job is requesting more than allowed "
"with assoc %u(%s/%s/%s) max tres(%s%s%s) "
@@ -2478,7 +2735,8 @@ extern bool acct_policy_job_runnable_post_select(
job_ptr->limit_set.tres,
1, 0, 1)) {
xfree(job_ptr->state_desc);
- job_ptr->state_reason = WAIT_ASSOC_MAX_CPUS_PER_JOB;
+ job_ptr->state_reason = get_tres_state_reason(
+ tres_pos, WAIT_ASSOC_MAX_UNK_PER_JOB);
debug2("job %u is being held, "
"the job is requesting more than allowed "
"with assoc %u(%s/%s/%s) max tres(%s%s%s) "
@@ -2567,7 +2825,7 @@ extern uint32_t acct_policy_get_max_nodes(struct job_record *job_ptr,
if (grp_nodes < max_nodes_limit) {
max_nodes_limit = grp_nodes;
- *wait_reason = WAIT_QOS_GRP_NODES;
+ *wait_reason = WAIT_QOS_GRP_NODE;
}
}
@@ -2581,7 +2839,7 @@ extern uint32_t acct_policy_get_max_nodes(struct job_record *job_ptr,
max_nodes_limit)) {
max_nodes_limit =
assoc_ptr->grp_tres_ctld[TRES_ARRAY_NODE];
- *wait_reason = WAIT_ASSOC_GRP_NODES;
+ *wait_reason = WAIT_ASSOC_GRP_NODE;
grp_set = 1;
}
--
GitLab