From 49d483dbae704047f876f3b2c840eafe7f1f2279 Mon Sep 17 00:00:00 2001
From: Morris Jette <jette@schedmd.com>
Date: Thu, 14 Apr 2016 09:39:27 -0700
Subject: [PATCH] Set burst buffer reason for job
If a job fails stage in, set its reason to BurstBufferOperation
with a string describing what happened. Previously the reason was
set to AdminHeld on stage-in failure.
---
src/plugins/burst_buffer/cray/burst_buffer_cray.c | 2 +-
src/slurmctld/job_mgr.c | 1 +
src/slurmctld/job_scheduler.c | 1 +
src/slurmctld/node_scheduler.c | 1 +
4 files changed, 4 insertions(+), 1 deletion(-)
diff --git a/src/plugins/burst_buffer/cray/burst_buffer_cray.c b/src/plugins/burst_buffer/cray/burst_buffer_cray.c
index 25605301121..151867b05cc 100644
--- a/src/plugins/burst_buffer/cray/burst_buffer_cray.c
+++ b/src/plugins/burst_buffer/cray/burst_buffer_cray.c
@@ -3231,7 +3231,7 @@ static void _kill_job(struct job_record *job_ptr)
job_ptr->priority = 0; /* Hold job */
build_cg_bitmap(job_ptr);
job_ptr->exit_code = 1;
- job_ptr->state_reason = WAIT_HELD;
+ job_ptr->state_reason = FAIL_BURST_BUFFER_OP;
xfree(job_ptr->state_desc);
job_ptr->state_desc = xstrdup("Burst buffer pre_run error");
job_completion_logger(job_ptr, false);
diff --git a/src/slurmctld/job_mgr.c b/src/slurmctld/job_mgr.c
index 5e599a28d3f..4e00fefe687 100644
--- a/src/slurmctld/job_mgr.c
+++ b/src/slurmctld/job_mgr.c
@@ -9570,6 +9570,7 @@ static bool _top_priority(struct job_record *job_ptr)
if ((!top) && detail_ptr) { /* not top prio */
if (job_ptr->priority == 0) { /* user/admin hold */
if (job_ptr->state_reason != FAIL_BAD_CONSTRAINTS
+ && (job_ptr->state_reason != FAIL_BURST_BUFFER_OP)
&& (job_ptr->state_reason != WAIT_HELD)
&& (job_ptr->state_reason != WAIT_HELD_USER)
&& job_ptr->state_reason != WAIT_MAX_REQUEUE) {
diff --git a/src/slurmctld/job_scheduler.c b/src/slurmctld/job_scheduler.c
index b7b12c96d8a..12997e8dfec 100644
--- a/src/slurmctld/job_scheduler.c
+++ b/src/slurmctld/job_scheduler.c
@@ -225,6 +225,7 @@ static bool _job_runnable_test1(struct job_record *job_ptr, bool sched_plugin)
job_ptr->start_time = (time_t) 0;
if (job_ptr->priority == 0) { /* held */
if (job_ptr->state_reason != FAIL_BAD_CONSTRAINTS
+ && (job_ptr->state_reason != FAIL_BURST_BUFFER_OP)
&& (job_ptr->state_reason != WAIT_HELD)
&& (job_ptr->state_reason != WAIT_HELD_USER)
&& job_ptr->state_reason != WAIT_MAX_REQUEUE) {
diff --git a/src/slurmctld/node_scheduler.c b/src/slurmctld/node_scheduler.c
index d81393526b5..026de5ff337 100644
--- a/src/slurmctld/node_scheduler.c
+++ b/src/slurmctld/node_scheduler.c
@@ -1855,6 +1855,7 @@ extern int select_nodes(struct job_record *job_ptr, bool test_only,
if (job_ptr->priority == 0) { /* user/admin hold */
if (job_ptr->state_reason != FAIL_BAD_CONSTRAINTS
+ && (job_ptr->state_reason != FAIL_BURST_BUFFER_OP)
&& (job_ptr->state_reason != WAIT_HELD)
&& (job_ptr->state_reason != WAIT_HELD_USER)
&& job_ptr->state_reason != WAIT_MAX_REQUEUE) {
--
GitLab