Fix: increase timings for osd release

Signed-off-by: Peter Sabaini <[email protected]>

.github/workflows/tests.yml

@@ -689,6 +689,10 @@ jobs:
         set -uex
         lxc exec node-wrk0 -- sh -c "sudo microceph status"
 
+    - name: Setup upterm session
+      if: ${{ failure() && runner.debug }}
+      uses: lhotari/action-upterm@v1
+
   cluster-tests:
     name: Test MicroCeph Cluster features.
     runs-on: ubuntu-22.04

microceph/ceph/start.go

@@ -40,12 +40,12 @@ func getCurrentVersion() (string, error) {
 }
 
 // checkVersions checks if all Ceph services are running the same version
-// retry up to 3 times if multiple versions are detected to allow for upgrades to complete as they are performed
+// retry up to 10 times if multiple versions are detected to allow for upgrades to complete as they are performed
 // concurrently
 func checkVersions() (bool, error) {
 	const (
-		maxRetries = 3
-		retryDelay = 5 * time.Second
+		maxRetries = 10
+		retryDelay = 10 * time.Second
 	)
 
 	for attempt := 0; attempt < maxRetries; attempt++ {
@@ -202,6 +202,7 @@ func Start(ctx context.Context, s interfaces.StateInterface) error {
 	}()
 
 	go func() {
+		time.Sleep(10 * time.Second) // wait for the mons to converge
 		err := PostRefresh()
 		if err != nil {
 			logger.Errorf("PostRefresh failed: %v", err)

microceph/ceph/start_test.go

@@ -85,7 +85,7 @@ func (s *startSuite) TestMultipleVersionsPresent() {
 }`
 
 	r.On("RunCommand", "ceph", "-v").Return(version, nil).Once()
-	r.On("RunCommand", "ceph", "versions").Return(versionsJson, nil).Times(3)
+	r.On("RunCommand", "ceph", "versions").Return(versionsJson, nil).Times(10)
 	processExec = r
 
 	err := PostRefresh()