From e9643c61392e04ef5b8a67a613a1334659c58e98 Mon Sep 17 00:00:00 2001
From: rawdaGastan <rody.rawda.rr@gmail.com>
Date: Wed, 7 Aug 2024 17:11:29 +0300
Subject: [PATCH] result an error not deleting workloads after a crash

---
 client/node.go          |  2 +-
 pkg/container/watch.go  |  4 ++--
 pkg/provision/engine.go | 17 ++++++-----------
 pkg/vm/monitor.go       |  5 ++---
 4 files changed, 11 insertions(+), 17 deletions(-)

diff --git a/client/node.go b/client/node.go
index 073d0154b..72b6dea33 100644
--- a/client/node.go
+++ b/client/node.go
@@ -154,7 +154,7 @@ func (n *NodeClient) DeploymentList(ctx context.Context) (dls []gridtypes.Deploy
 	return
 }
 
-// DeploymentGet gets a deployment via contract ID
+// DeploymentChanges gets a deployment changes via contract ID
 func (n *NodeClient) DeploymentChanges(ctx context.Context, contractID uint64) (changes []gridtypes.Workload, err error) {
 	const cmd = "zos.deployment.changes"
 	in := args{
diff --git a/pkg/container/watch.go b/pkg/container/watch.go
index 157ca8829..ed6906cc6 100644
--- a/pkg/container/watch.go
+++ b/pkg/container/watch.go
@@ -55,11 +55,11 @@ func (c *Module) handlerEventTaskExit(ctx context.Context, ns string, event *eve
 		<-time.After(restartDelay) // wait for 2 seconds
 		reason = c.start(ns, event.ContainerID)
 	} else {
-		reason = fmt.Errorf("deleting container due to so many crashes")
+		reason = fmt.Errorf("container failed due to so many crashes")
 	}
 
 	if reason != nil {
-		log.Debug().Err(reason).Msg("deleting container due to restart error")
+		log.Debug().Err(reason).Msg("resulting error for container due to restart error")
 
 		stub := stubs.NewProvisionStub(c.client)
 		if err := stub.DecommissionCached(ctx, event.ContainerID, reason.Error()); err != nil {
diff --git a/pkg/provision/engine.go b/pkg/provision/engine.go
index 5880fa27e..0135544d5 100644
--- a/pkg/provision/engine.go
+++ b/pkg/provision/engine.go
@@ -989,18 +989,13 @@ func (e *NativeEngine) DecommissionCached(id string, reason string) error {
 		return nil
 	}
 
-	//to bad we have to repeat this here
-	ctx := context.WithValue(context.Background(), engineKey{}, e)
-	ctx = withDeployment(ctx, twin, dlID)
-
-	ctx, cancel := context.WithTimeout(ctx, 3*time.Minute)
-	defer cancel()
-
-	err = e.uninstallWorkload(ctx, &gridtypes.WorkloadWithID{Workload: &wl, ID: globalID},
-		fmt.Sprintf("workload decommissioned by system, reason: %s", reason),
-	)
+	result := gridtypes.Result{
+		State:   gridtypes.StateError,
+		Error:   reason,
+		Created: gridtypes.Timestamp(time.Now().Unix()),
+	}
 
-	return err
+	return e.storage.Transaction(twin, dlID, wl.WithResults(result))
 }
 
 func (n *NativeEngine) CreateOrUpdate(twin uint32, deployment gridtypes.Deployment, update bool) error {
diff --git a/pkg/vm/monitor.go b/pkg/vm/monitor.go
index 9a068afcd..594060bea 100644
--- a/pkg/vm/monitor.go
+++ b/pkg/vm/monitor.go
@@ -216,12 +216,11 @@ func (m *Module) monitorID(ctx context.Context, running map[string]Process, id s
 			reason = m.withLogs(m.logsPath(id), err)
 		}
 	} else {
-		reason = fmt.Errorf("deleting vm due to so many crashes")
+		reason = fmt.Errorf("resulting error for vm due to so many crashes")
 	}
 
 	if reason != nil {
-		log.Debug().Err(reason).Msg("deleting vm due to restart error")
-		m.removeConfig(id)
+		log.Debug().Err(reason).Msg("resulting error for vm due to restart error")
 
 		if err := stub.DecommissionCached(ctx, id, reason.Error()); err != nil {
 			return errors.Wrapf(err, "failed to decommission reservation '%s'", id)