fix: clean install fixes and pin mc-iam-manager v0.6.4 / mc-web-console v0.6.6

bin/installAll.sh

@@ -168,6 +168,35 @@ STARTUP_WAVES=(
     "mc-application-manager mc-workflow-manager mc-cost-optimizer-fe"
 )
 
+# Starts every wave detached (-d), then one final full-stack `run -d` for the
+# services no wave entry point reaches (cost-optimizer collectors/rightsizers,
+# cb-mapui, ...). Waves must always be detached: an attached `compose up` never
+# returns, so later waves would never start.
+run_startup_waves() {
+    local wave_num=0
+    local run_exit
+    for wave_services in "${STARTUP_WAVES[@]}"; do
+        wave_num=$((wave_num + 1))
+        echo ""
+        echo "---- Wave $wave_num/${#STARTUP_WAVES[@]}: $wave_services ----"
+        ./mcc infra run -d -s "$wave_services"
+        run_exit=$?
+        if [ $run_exit -ne 0 ]; then
+            report_run_failure "$run_exit" "Wave $wave_num ($wave_services)"
+            exit 1
+        fi
+    done
+
+    echo ""
+    echo "---- Remaining services (not reached by any wave) ----"
+    ./mcc infra run -d
+    run_exit=$?
+    if [ $run_exit -ne 0 ]; then
+        report_run_failure "$run_exit" "remaining services"
+        exit 1
+    fi
+}
+
 # Prints a consistent failure banner for a failed `./mcc infra run` invocation.
 report_run_failure() {
     local exit_code="$1"
@@ -543,20 +572,14 @@ case $RUN_MODE in
             exit 1
         fi
 
-        wave_num=0
-        for wave_services in "${STARTUP_WAVES[@]}"; do
-            wave_num=$((wave_num + 1))
-            echo ""
-            echo "---- Wave $wave_num/${#STARTUP_WAVES[@]}: $wave_services ----"
-            ./mcc infra run -s "$wave_services"
-            run_exit=$?
-            if [ $run_exit -ne 0 ]; then
-                report_run_failure "$run_exit" "Wave $wave_num ($wave_services)"
-                exit 1
-            fi
-        done
-
+        run_startup_waves
         check_post_initial
+
+        # Attach to the whole stack only after every wave is up (Ctrl+C to detach;
+        # containers keep running)
+        echo ""
+        echo "All services started. Attaching to logs (Ctrl+C to stop following)..."
+        ./mcc infra run
         ;;
     background)
         echo ""
@@ -579,18 +602,7 @@ case $RUN_MODE in
         echo "Image download and initial setup in progress..."
         echo ""
 
-        wave_num=0
-        for wave_services in "${STARTUP_WAVES[@]}"; do
-            wave_num=$((wave_num + 1))
-            echo ""
-            echo "---- Wave $wave_num/${#STARTUP_WAVES[@]}: $wave_services ----"
-            ./mcc infra run -d -s "$wave_services"
-            run_exit=$?
-            if [ $run_exit -ne 0 ]; then
-                report_run_failure "$run_exit" "Wave $wave_num ($wave_services)"
-                exit 1
-            fi
-        done
+        run_startup_waves
 
         echo ""
         echo "Image download and initial setup completed."

conf/api.yaml

@@ -11,7 +11,7 @@ services:
       password:
 
   mc-iam-manager:
-    version: 0.5.2
+    version: 0.6.4
     baseurl: http://mc-iam-manager:5000
     auth:
       type: bearer
@@ -25,7 +25,7 @@ services:
       password: spider
 
   mc-web-console:
-    version: 0.5.2
+    version: 0.6.6
     baseurl: http://mc-web-console:3000
     auth:
       type: bearer

conf/docker/.env.setup

@@ -159,7 +159,7 @@ MC_IAM_MANAGER_DEFAULT_WORKSPACE_NAME=ws01
 MC_IAM_MANAGER_AWS_STS_ENDPOINT=https://sts.amazonaws.com
 MC_IAM_MANAGER_AWS_ACCOUNT_ID=notyet
 MC_IAM_MANAGER_AWS_IDENTITY_PROVIDER_ARN=arn:aws:iam::${MC_IAM_MANAGER_AWS_ACCOUNT_ID}:oidc-provider/${MC_IAM_MANAGER_KEYCLOAK_DOMAIN}/realms/${MC_IAM_MANAGER_KEYCLOAK_OIDC_CLIENT_NAME}
-MC_IAM_MANAGER_AWS_IDENTITY_ROLE_ARN=arn:aws:iam::${MC_IAM_MANAGER_KEYCLOAK_DOMAIN}:role/mciam-platformadmin
+MC_IAM_MANAGER_AWS_IDENTITY_ROLE_ARN=arn:aws:iam::${MC_IAM_MANAGER_AWS_ACCOUNT_ID}:role/mciam-platformadmin
 
 
 MC_IAM_MANAGER_CSP_ROLE_PREFIX=mciam

conf/docker/api.yaml

@@ -11,7 +11,7 @@ services:
       password:
 
   mc-iam-manager:
-    version: 0.5.2
+    version: 0.6.4
     baseurl: http://mc-iam-manager:5000
     auth:
       type: bearer
@@ -25,7 +25,7 @@ services:
       password: spider
 
   mc-web-console:
-    version: 0.5.2
+    version: 0.6.6
     baseurl: http://mc-web-console:3000
     auth:
       type: bearer

conf/docker/conf/mc-web-console/api/conf/api.yaml

@@ -11,7 +11,7 @@ services:
       password:
 
   mc-iam-manager:
-    version: 0.5.1
+    version: 0.6.4
     baseurl: http://mc-iam-manager:5000
     auth:
       type: bearer
@@ -25,7 +25,7 @@ services:
       password: default
 
   mc-web-console:
-    version: 0.5.2
+    version: 0.6.6
     baseurl: http://mc-web-console-api:3000
     auth:
       type: bearer

conf/docker/docker-compose.mini.yaml

@@ -219,7 +219,7 @@ services:
 
   mc-iam-manager:
     container_name: mc-iam-manager
-    image: cloudbaristaorg/mc-iam-manager:0.6.3
+    image: cloudbaristaorg/mc-iam-manager:0.6.4
     restart: unless-stopped
     networks:
       - mc-iam-manager-network
@@ -387,7 +387,7 @@ services:
             exec docker-entrypoint.sh postgres"
 
   mc-web-console-api:
-    image: csescsta/mc-web-console-api:edge
+    image: cloudbaristaorg/mc-web-console-api:0.6.6
     pull_policy: always
     container_name: mc-web-console-api
     platform: linux/amd64
@@ -436,7 +436,7 @@ services:
       <<: *default-health-check
 
   mc-web-console-front:
-    image: csescsta/mc-web-console-front:edge
+    image: cloudbaristaorg/mc-web-console-front:0.6.6
     pull_policy: always
     container_name: mc-web-console-front
     platform: linux/amd64

conf/docker/docker-compose.yaml

@@ -277,7 +277,7 @@ services:
 
   mc-iam-manager:
     container_name: mc-iam-manager
-    image: cloudbaristaorg/mc-iam-manager:0.6.3
+    image: cloudbaristaorg/mc-iam-manager:0.6.4
     restart: unless-stopped
     networks:
       - mc-iam-manager-network
@@ -1042,7 +1042,10 @@ services:
           curl -fsS -u "$${JENKINS_USERNAME}:$${JENKINS_PASSWORD}"
           http://localhost:8080/api/json > /dev/null
       <<: *default-health-check
-      start_period: 6m
+      # First boot downloads ~100 plugins and restarts Jenkins; a clean install
+      # was observed taking ~9.5 min. start_period only suppresses failures, so a
+      # generous value costs nothing once Jenkins is actually ready.
+      start_period: 15m
 
 
   mc-workflow-manager:
@@ -1176,7 +1179,7 @@ services:
             exec docker-entrypoint.sh postgres"
 
   mc-web-console-api:
-    image: cloudbaristaorg/mc-web-console-api:0.6.5
+    image: cloudbaristaorg/mc-web-console-api:0.6.6
     pull_policy: always
     container_name: mc-web-console-api
     platform: linux/amd64
@@ -1220,7 +1223,7 @@ services:
       <<: *default-health-check
 
   mc-web-console-front:
-    image: cloudbaristaorg/mc-web-console-front:0.6.5
+    image: cloudbaristaorg/mc-web-console-front:0.6.6
     pull_policy: always
     container_name: mc-web-console-front
     platform: linux/amd64

conf/docker/tool/init.groovy.d/basic-security.groovy

@@ -97,6 +97,37 @@ plugins.each { pluginName ->
     }
 }
 
+// deploy() also queues each dependency as its own InstallationJob, and a failed
+// download (e.g. "Connection reset" from the update site) does not throw -- it
+// only leaves that job in Failure state. Without a retry the missing dependency
+// (e.g. joda-time-api) makes dependents fail to load after restart, and Jenkins
+// only converges after several restart cycles.
+def failedInstalls = {
+    def latestJobs = [:]
+    uc.getJobs().findAll { it instanceof UpdateCenter.InstallationJob }.each { job ->
+        latestJobs[job.plugin.name] = job
+    }
+    latestJobs.findAll { name, job -> job.status instanceof UpdateCenter.DownloadJob.Failure }.keySet()
+}
+
+def maxInstallRetries = 5
+for (int attempt = 1; attempt <= maxInstallRetries; attempt++) {
+    def failed = failedInstalls()
+    if (failed.isEmpty()) {
+        break
+    }
+    println "--> Retrying failed plugin downloads (${attempt}/${maxInstallRetries}): ${failed.join(', ')}"
+    sleep(5000L * attempt)
+    failed.each { pluginName ->
+        uc.getPlugin(pluginName)?.deploy()?.get()
+    }
+}
+
+def stillFailed = failedInstalls()
+if (!stillFailed.isEmpty()) {
+    throw new IllegalStateException("Jenkins plugin download failed after ${maxInstallRetries} retries: ${stillFailed.join(', ')}")
+}
+
 println "--> Saving Jenkins state..."
 instance.save()