[ARVADOS] created: b887b0b6185d0c4f51db511b54306686ac17f6ce

[Git user]

        at  b887b0b6185d0c4f51db511b54306686ac17f6ce (commit)


commit b887b0b6185d0c4f51db511b54306686ac17f6ce
Author: Peter Amstutz <peter.amstutz at curoverse.com>
Date:   Thu Dec 1 12:56:17 2016 -0500

    10649: Make errors emitted by squeue and scancel show up in logs.

diff --git a/services/crunch-dispatch-slurm/crunch-dispatch-slurm.go b/services/crunch-dispatch-slurm/crunch-dispatch-slurm.go
index 3c4f281..2e5908c 100644
--- a/services/crunch-dispatch-slurm/crunch-dispatch-slurm.go
+++ b/services/crunch-dispatch-slurm/crunch-dispatch-slurm.go
@@ -302,12 +302,13 @@ func run(dispatcher *dispatch.Dispatcher,
 
 				// Mutex between squeue sync and running sbatch or scancel.
 				squeueUpdater.SlurmLock.Lock()
-				err := scancelCmd(container).Run()
+				cmd := scancelCmd(container)
+				msg, err := cmd.CombinedOutput()
 				squeueUpdater.SlurmLock.Unlock()
 
 				if err != nil {
-					log.Printf("Error stopping container %s with scancel: %v",
-						container.UUID, err)
+					log.Printf("Error stopping container %s with %v %v: %v %v",
+						container.UUID, cmd.Path, cmd.Args, err, string(msg))
 					if squeueUpdater.CheckSqueue(container.UUID) {
 						log.Printf("Container %s is still in squeue after scancel.",
 							container.UUID)
diff --git a/services/crunch-dispatch-slurm/squeue.go b/services/crunch-dispatch-slurm/squeue.go
index 61decde..1f5d586 100644
--- a/services/crunch-dispatch-slurm/squeue.go
+++ b/services/crunch-dispatch-slurm/squeue.go
@@ -2,6 +2,7 @@ package main
 
 import (
 	"bufio"
+	"io/ioutil"
 	"log"
 	"os/exec"
 	"sync"
@@ -45,31 +46,47 @@ func (squeue *Squeue) RunSqueue() {
 		log.Printf("Error creating stdout pipe for squeue: %v", err)
 		return
 	}
+
+	stderrReader, err := cmd.StderrPipe()
+	if err != nil {
+		log.Printf("Error creating stderr pipe for squeue: %v", err)
+		return
+	}
+
 	err = cmd.Start()
 	if err != nil {
 		log.Printf("Error running squeue: %v", err)
 		return
 	}
+
+	stderrChan := make(chan []byte)
+	go func() {
+		b, _ := ioutil.ReadAll(stderrReader)
+		stderrReader.Close()
+		stderrChan <- b
+	}()
+
 	scanner := bufio.NewScanner(sq)
 	for scanner.Scan() {
 		newSqueueContents = append(newSqueueContents, scanner.Text())
 	}
-	if err := scanner.Err(); err != nil {
-		cmd.Wait()
-		log.Printf("Error reading from squeue pipe: %v", err)
-		return
-	}
 
 	err = cmd.Wait()
+	stderrmsg := <-stderrChan
+
+	if scanner.Err() != nil {
+		log.Printf("Error reading from squeue pipe: %v", err)
+	}
 	if err != nil {
-		log.Printf("Error running squeue: %v", err)
-		return
+		log.Printf("Error running %v %v: %v %q", cmd.Path, cmd.Args, err, string(stderrmsg))
 	}
 
-	squeue.squeueCond.L.Lock()
-	squeue.squeueContents = newSqueueContents
-	squeue.squeueCond.Broadcast()
-	squeue.squeueCond.L.Unlock()
+	if scanner.Err() == nil && err == nil {
+		squeue.squeueCond.L.Lock()
+		squeue.squeueContents = newSqueueContents
+		squeue.squeueCond.Broadcast()
+		squeue.squeueCond.L.Unlock()
+	}
 }
 
 // CheckSqueue checks if a given container UUID is in the slurm queue.  This

-----------------------------------------------------------------------


hooks/post-receive
--