From 56abc00b78176abf2d47c7a48df88d48609a51c4 Mon Sep 17 00:00:00 2001
From: Jeff Layton <jlayton@redhat.com>
Date: Wed, 27 Feb 2019 10:43:30 -0500
Subject: [PATCH 1/2] ceph: fix invalidArgErrorCode definition

We should probably not hardcode this value here, and instead use
whatever the syscall module thinks this is.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
 pkg/operator/ceph/cluster/mgr/dashboard.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pkg/operator/ceph/cluster/mgr/dashboard.go b/pkg/operator/ceph/cluster/mgr/dashboard.go
index 69a1da5ec..29d3592c4 100755
--- a/pkg/operator/ceph/cluster/mgr/dashboard.go
+++ b/pkg/operator/ceph/cluster/mgr/dashboard.go
@@ -42,7 +42,7 @@ const (
 	passwordLength                 = 10
 	passwordKeyName                = "password"
 	certAlreadyConfiguredErrorCode = 5
-	invalidArgErrorCode            = 22
+	invalidArgErrorCode            = int(syscall.EINVAL)
 )
 
 var (

From c0910dab2cb5a15c0a070d7c9729e216aafa6f66 Mon Sep 17 00:00:00 2001
From: Jeff Layton <jlayton@redhat.com>
Date: Wed, 27 Feb 2019 10:12:42 -0500
Subject: [PATCH 2/2] ceph: retry "ceph orchestrator set backend rook" if
 command returns EINVAL

Use similar method as that used by the dashboard initialization code.

Fixes: https://github.com/rook/rook/issues/2617
Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
 pkg/operator/ceph/cluster/mgr/orchestrator.go | 22 +++++++++++++++++--
 1 file changed, 20 insertions(+), 2 deletions(-)

diff --git a/pkg/operator/ceph/cluster/mgr/orchestrator.go b/pkg/operator/ceph/cluster/mgr/orchestrator.go
index 0319ab9e0..a1d1e0471 100644
--- a/pkg/operator/ceph/cluster/mgr/orchestrator.go
+++ b/pkg/operator/ceph/cluster/mgr/orchestrator.go
@@ -19,6 +19,7 @@ package mgr
 
 import (
 	"fmt"
+	"time"
 
 	cephv1 "github.com/rook/rook/pkg/apis/ceph.rook.io/v1"
 	"github.com/rook/rook/pkg/daemon/ceph/client"
@@ -29,6 +30,10 @@ const (
 	rookModuleName         = "rook"
 )
 
+var (
+	orchestratorInitWaitTime = 5 * time.Second
+)
+
 // Ceph docs about the orchestrator modules: http://docs.ceph.com/docs/master/mgr/orchestrator_cli/
 func (c *Cluster) configureOrchestratorModules() error {
 	if !cephv1.VersionAtLeast(c.cephVersion.Name, cephv1.Nautilus) {
@@ -42,8 +47,21 @@ func (c *Cluster) configureOrchestratorModules() error {
 	if err := client.MgrEnableModule(c.context, c.Namespace, rookModuleName, true); err != nil {
 		return fmt.Errorf("failed to enable mgr rook module. %+v", err)
 	}
-	if _, err := client.ExecuteCephCommand(c.context, c.Namespace, []string{"orchestrator", "set", "backend", "rook"}); err != nil {
-		return fmt.Errorf("failed to set rook as the orchestrator backend. %+v", err)
+	// retry a few times in the case that the mgr module is not ready to accept commands
+	for i := 0; i < 5; i++ {
+		_, err := client.ExecuteCephCommand(c.context, c.Namespace, []string{"orchestrator", "set", "backend", "rook"})
+		if err != nil {
+			exitCode, parsed := c.exitCode(err)
+			if parsed {
+				if exitCode == invalidArgErrorCode {
+					logger.Infof("orchestrator module is not ready yet. trying again...")
+					time.Sleep(orchestratorInitWaitTime)
+					continue
+				}
+			}
+			return fmt.Errorf("failed to set rook as the orchestrator backend. %+v", err)
+		}
+		break
 	}
 
 	return nil