X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/72a8b3582d925ea30fe78697ff76bafb20d8bd9e..6d8326a0f1521b0d7c49cd90dc775d409906bc49:/services/crunch-dispatch-slurm/crunch-dispatch-slurm.go

diff --git a/services/crunch-dispatch-slurm/crunch-dispatch-slurm.go b/services/crunch-dispatch-slurm/crunch-dispatch-slurm.go
index 75e6146f53..584db38edf 100644
--- a/services/crunch-dispatch-slurm/crunch-dispatch-slurm.go
+++ b/services/crunch-dispatch-slurm/crunch-dispatch-slurm.go
@@ -7,7 +7,6 @@ package main
 // Dispatcher service for Crunch that submits containers to the slurm queue.
 
 import (
-	"bytes"
 	"context"
 	"flag"
 	"fmt"
@@ -18,14 +17,14 @@ import (
 	"strings"
 	"time"
 
-	"git.curoverse.com/arvados.git/lib/config"
-	"git.curoverse.com/arvados.git/lib/dispatchcloud"
-	"git.curoverse.com/arvados.git/sdk/go/arvados"
-	"git.curoverse.com/arvados.git/sdk/go/arvadosclient"
-	"git.curoverse.com/arvados.git/sdk/go/dispatch"
+	"git.arvados.org/arvados.git/lib/config"
+	"git.arvados.org/arvados.git/lib/dispatchcloud"
+	"git.arvados.org/arvados.git/sdk/go/arvados"
+	"git.arvados.org/arvados.git/sdk/go/arvadosclient"
+	"git.arvados.org/arvados.git/sdk/go/dispatch"
 	"github.com/coreos/go-systemd/daemon"
+	"github.com/ghodss/yaml"
 	"github.com/sirupsen/logrus"
-	"gopkg.in/yaml.v2"
 )
 
 type logger interface {
@@ -109,7 +108,6 @@ func (disp *Dispatcher) configure(prog string, args []string) error {
 
 	disp.logger.Printf("crunch-dispatch-slurm %s started", version)
 
-	loader.LegacyComponentConfig = loader.CrunchDispatchSlurmPath
 	cfg, err := loader.Load()
 	if err != nil {
 		return err
@@ -133,8 +131,10 @@ func (disp *Dispatcher) configure(prog string, args []string) error {
 		if disp.Client.Insecure {
 			os.Setenv("ARVADOS_API_HOST_INSECURE", "1")
 		}
-		os.Setenv("ARVADOS_KEEP_SERVICES", strings.Join(disp.Client.KeepServiceURIs, " "))
 		os.Setenv("ARVADOS_EXTERNAL_CLIENT", "")
+		for k, v := range disp.cluster.Containers.SLURM.SbatchEnvironmentVariables {
+			os.Setenv(k, v)
+		}
 	} else {
 		disp.logger.Warnf("Client credentials missing from config, so falling back on environment variables (deprecated).")
 	}
@@ -201,7 +201,7 @@ var containerUuidPattern = regexp.MustCompile(`^[a-z0-9]{5}-dz642-[a-z0-9]{15}$`
 // Cancelled or Complete. See https://dev.arvados.org/issues/10979
 func (disp *Dispatcher) checkSqueueForOrphans() {
 	for _, uuid := range disp.sqCheck.All() {
-		if !containerUuidPattern.MatchString(uuid) {
+		if !containerUuidPattern.MatchString(uuid) || !strings.HasPrefix(uuid, disp.cluster.ClusterID) {
 			continue
 		}
 		err := disp.TrackContainer(uuid)
@@ -254,6 +254,7 @@ func (disp *Dispatcher) submit(container arvados.Container, crunchRunCommand []s
 	// append() here avoids modifying crunchRunCommand's
 	// underlying array, which is shared with other goroutines.
 	crArgs := append([]string(nil), crunchRunCommand...)
+	crArgs = append(crArgs, "--runtime-engine="+disp.cluster.Containers.RuntimeEngine)
 	crArgs = append(crArgs, container.UUID)
 	crScript := strings.NewReader(execScript(crArgs))
 
@@ -269,7 +270,7 @@ func (disp *Dispatcher) submit(container arvados.Container, crunchRunCommand []s
 // already in the queue).  Cancel the slurm job if the container's
 // priority changes to zero or its state indicates it's no longer
 // running.
-func (disp *Dispatcher) runContainer(_ *dispatch.Dispatcher, ctr arvados.Container, status <-chan arvados.Container) {
+func (disp *Dispatcher) runContainer(_ *dispatch.Dispatcher, ctr arvados.Container, status <-chan arvados.Container) error {
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()
 
@@ -277,37 +278,9 @@ func (disp *Dispatcher) runContainer(_ *dispatch.Dispatcher, ctr arvados.Contain
 		log.Printf("Submitting container %s to slurm", ctr.UUID)
 		cmd := []string{disp.cluster.Containers.CrunchRunCommand}
 		cmd = append(cmd, disp.cluster.Containers.CrunchRunArgumentsList...)
-		if err := disp.submit(ctr, cmd); err != nil {
-			var text string
-			if err, ok := err.(dispatchcloud.ConstraintsNotSatisfiableError); ok {
-				var logBuf bytes.Buffer
-				fmt.Fprintf(&logBuf, "cannot run container %s: %s\n", ctr.UUID, err)
-				if len(err.AvailableTypes) == 0 {
-					fmt.Fprint(&logBuf, "No instance types are configured.\n")
-				} else {
-					fmt.Fprint(&logBuf, "Available instance types:\n")
-					for _, t := range err.AvailableTypes {
-						fmt.Fprintf(&logBuf,
-							"Type %q: %d VCPUs, %d RAM, %d Scratch, %f Price\n",
-							t.Name, t.VCPUs, t.RAM, t.Scratch, t.Price,
-						)
-					}
-				}
-				text = logBuf.String()
-				disp.UpdateState(ctr.UUID, dispatch.Cancelled)
-			} else {
-				text = fmt.Sprintf("Error submitting container %s to slurm: %s", ctr.UUID, err)
-			}
-			log.Print(text)
-
-			lr := arvadosclient.Dict{"log": arvadosclient.Dict{
-				"object_uuid": ctr.UUID,
-				"event_type":  "dispatch",
-				"properties":  map[string]string{"text": text}}}
-			disp.Arv.Create("logs", lr, nil)
-
-			disp.Unlock(ctr.UUID)
-			return
+		err := disp.submit(ctr, cmd)
+		if err != nil {
+			return err
 		}
 	}
 
@@ -336,7 +309,7 @@ func (disp *Dispatcher) runContainer(_ *dispatch.Dispatcher, ctr arvados.Contain
 			case dispatch.Locked:
 				disp.Unlock(ctr.UUID)
 			}
-			return
+			return nil
 		case updated, ok := <-status:
 			if !ok {
 				log.Printf("container %s is done: cancel slurm job", ctr.UUID)