[crmsh-4.6] Setup bootstrap stages dependency

crmsh/bootstrap.py

@@ -47,6 +47,7 @@
 from .sh import ShellUtils
 from .ui_node import NodeMgmt
 from .user_of_host import UserOfHost, UserNotFoundError
+import crmsh.healthcheck
 
 logger = log.setup_logger(__name__)
 logger_utils = log.LoggerUtils(logger)
@@ -74,7 +75,11 @@
         "/etc/drbd.conf", "/etc/drbd.d", "/etc/ha.d/ldirectord.cf", "/etc/lvm/lvm.conf", "/etc/multipath.conf",
         "/etc/samba/smb.conf", SYSCONFIG_NFS, SYSCONFIG_PCMK, SYSCONFIG_SBD, PCMK_REMOTE_AUTH, WATCHDOG_CFG,
         PROFILES_FILE, CRM_CFG, SBD_SYSTEMD_DELAY_START_DIR)
-INIT_STAGES = ("ssh", "csync2", "csync2_remote", "qnetd_remote", "corosync", "remote_auth", "sbd", "cluster", "ocfs2", "admin", "qdevice")
+
+INIT_STAGES_EXTERNAL = ("ssh", "csync2", "corosync", "sbd", "cluster", "ocfs2", "admin", "qdevice")
+INIT_STAGES_INTERNAL = ("csync2_remote", "qnetd_remote", "remote_auth")
+INIT_STAGES_ALL = INIT_STAGES_EXTERNAL + INIT_STAGES_INTERNAL
+JOIN_STAGES_EXTERNAL = ("ssh", "csync2", "ssh_merge", "cluster")
 
 
 class Context(object):
@@ -232,15 +237,40 @@
         """
         Validate cluster_node on join side
         """
-        if self.cluster_node and self.type == 'join':
+        if self.type == "join" and self.cluster_node:
             user, node = _parse_user_at_host(self.cluster_node, None)
             try:
                 # self.cluster_node might be hostname or IP address
                 ip_addr = socket.gethostbyname(node)
                 if utils.InterfacesInfo.ip_in_local(ip_addr):
                     utils.fatal("Please specify peer node's hostname or IP address")
             except socket.gaierror as err:
-                utils.fatal("\"{}\": {}".format(node, err))
+                utils.fatal(f"\"{node}\": {err}")
+
+    def _validate_stage(self):
+        """
+        Validate stage argument
+        """
+        if not self.stage:
+            if self.cluster_is_running:
+                utils.fatal("Cluster is already running!")
+            return
+
+        if self.type == "init":
+            if self.stage not in INIT_STAGES_ALL:
+                utils.fatal(f"Invalid stage: {self.stage}(available stages: {', '.join(INIT_STAGES_EXTERNAL)})")
+            if self.stage in ("admin", "qdevice", "ocfs2") and not self.cluster_is_running:
+                utils.fatal(f"Cluster is inactive, can't run '{self.stage}' stage")
+            if self.stage in ("corosync", "cluster") and self.cluster_is_running:
+                utils.fatal(f"Cluster is active, can't run '{self.stage}' stage")
+
+        elif self.type == "join":
+            if self.stage not in JOIN_STAGES_EXTERNAL:
+                utils.fatal(f"Invalid stage: {self.stage}(available stages: {', '.join(JOIN_STAGES_EXTERNAL)})")
+            if self.stage and self.cluster_node is None:
+                utils.fatal(f"Can't use stage({self.stage}) without specifying cluster node")
+            if self.stage in ("cluster", ) and self.cluster_is_running:
+                utils.fatal(f"Cluster is active, can't run '{self.stage}' stage")
 
     def validate_option(self):
         """
@@ -263,6 +293,7 @@
             self.skip_csync2 = utils.get_boolean(os.getenv("SKIP_CSYNC2_SYNC"))
         if self.skip_csync2 and self.stage:
             utils.fatal("-x option or SKIP_CSYNC2_SYNC can't be used with any stage")
+        self._validate_stage()
         self._validate_cluster_node()
         self._validate_nodes_option()
         self._validate_sbd_option()
@@ -553,7 +584,7 @@
         return False
 
 
-def check_prereqs(stage):
+def check_prereqs():
     warned = False
 
     if not my_hostname_resolves():
@@ -1710,6 +1741,9 @@
     change_user_shell('hacluster')
     swap_public_ssh_key_for_secondary_user(sh.cluster_shell(), seed_host, 'hacluster')
 
+    if _context.stage:
+        setup_passwordless_with_other_nodes(seed_host, seed_user)
+
 
 def join_ssh_with_ssh_agent(
         local_shell: sh.LocalShell,
@@ -2367,46 +2401,87 @@
     corosync.set_value("quorum.expected_votes", str(new_quorum))
 
 
+def ssh_stage_finished():
+    """
+    Dectect if the ssh stage is finished
+    """
+    feature_check = crmsh.healthcheck.PasswordlessHaclusterAuthenticationFeature()
+    return feature_check.check_quick() and feature_check.check_local([utils.this_node()])
+
+
+def csync2_stage_finished():
+    """
+    Dectect if the csync2 stage is finished
+    """
+    return ServiceManager().service_is_active(CSYNC2_SERVICE)
+
+
+def corosync_stage_finished():
+    """
+    Dectect if the corosync stage is finished
+    """
+    return os.path.exists(corosync.conf())
+
+
+INIT_STAGE_CHECKER = {
+        # stage: (function, is_internal)
+        "ssh": (ssh_stage_finished, False),
+        "csync2": (csync2_stage_finished, False),
+        "corosync": (corosync_stage_finished, False),
+        "remote_auth": (init_remote_auth, True),
+        "sbd": (lambda: True, False),
+        "upgradeutil": (init_upgradeutil, True),
+        "cluster": (is_online, False)
+}
+
+
+JOIN_STAGE_CHECKER = {
+        # stage: (function, is_internal)
+        "ssh": (ssh_stage_finished, False),
+        "csync2": (csync2_stage_finished, False),
+        "ssh_merge": (lambda: True, False),
+        "cluster": (is_online, False)
+}
+
+
+def check_stage_dependency(stage):
+    stage_checker = INIT_STAGE_CHECKER if _context.type == "init" else JOIN_STAGE_CHECKER
+    if stage not in stage_checker:
+        return
+    stage_order = list(stage_checker.keys())
+    for stage_name in stage_order:
+        if stage == stage_name:
+            break
+        func, is_internal = stage_checker[stage_name]
+        if is_internal:
+            func()
+        elif not func():
+            utils.fatal(f"Please run '{stage_name}' stage first")
+
+
 def bootstrap_init(context):
     """
     Init cluster process
     """
     global _context
     _context = context
+    stage = _context.stage
 
     init()
 
-    stage = _context.stage
-    if stage is None:
-        stage = ""
-
-    # vgfs stage requires running cluster, everything else requires inactive cluster,
-    # except ssh and csync2 (which don't care) and csync2_remote (which mustn't care,
-    # just in case this breaks ha-cluster-join on another node).
-    if stage in ("vgfs", "admin", "qdevice", "ocfs2"):
-        if not _context.cluster_is_running:
-            utils.fatal("Cluster is inactive - can't run %s stage" % (stage))
-    elif stage == "":
-        if _context.cluster_is_running:
-            utils.fatal("Cluster is currently active - can't run")
-    elif stage not in ("ssh", "csync2", "csync2_remote", "qnetd_remote", "sbd", "ocfs2"):
-        if _context.cluster_is_running:
-            utils.fatal("Cluster is currently active - can't run %s stage" % (stage))
-
     _context.load_profiles()
     _context.init_sbd_manager()
 
-    # Need hostname resolution to work, want NTP (but don't block csync2_remote)
-    if stage not in ('csync2_remote', 'qnetd_remote'):
-        check_tty()
-        if not check_prereqs(stage):
-            return
-    else:
+    if stage in ('csync2_remote', 'qnetd_remote'):
         args = _context.args
-        logger_utils.log_only_to_file("args: {}".format(args))
+        logger_utils.log_only_to_file(f"args: {args}")
         if len(args) != 2:
-            utils.fatal(f"Expected NODE argument to {stage} stage")
+            utils.fatal(f"Expected NODE argument for '{stage}' stage")
         _context.cluster_node = args[1]
+    else:
+        check_tty()
+        if not check_prereqs():
+            return
 
     if stage and _context.cluster_is_running and \
             not ServiceManager(shell=sh.ClusterShellAdaptorForLocalShell(sh.LocalShell())).service_is_active(CSYNC2_SERVICE):
@@ -2416,6 +2491,7 @@
         _context.node_list_in_cluster = [utils.this_node()]
 
     if stage != "":
+        check_stage_dependency(stage)
         globals()["init_" + stage]()
     else:
         init_ssh()
@@ -2492,15 +2568,13 @@
 
     check_tty()
 
-    corosync_active = ServiceManager(sh.ClusterShellAdaptorForLocalShell(sh.LocalShell())).service_is_active("corosync.service")
-    if corosync_active and _context.stage != "ssh":
-        utils.fatal("Abort: Cluster is currently active. Run this command on a node joining the cluster.")
-
-    if not check_prereqs("join"):
+    if not check_prereqs():
         return
 
     if _context.stage != "":
         remote_user, cluster_node = _parse_user_at_host(_context.cluster_node, _context.current_user)
+        init_upgradeutil()
+        check_stage_dependency(_context.stage)
         globals()["join_" + _context.stage](cluster_node, remote_user)
     else:
         if not _context.yes_to_all and _context.cluster_node is None:
@@ -2527,7 +2601,6 @@
                 service_manager = ServiceManager()
                 _context.node_list_in_cluster = utils.fetch_cluster_node_list_from_node(cluster_node)
                 setup_passwordless_with_other_nodes(cluster_node, remote_user)
-                join_remote_auth(cluster_node, remote_user)
                 _context.skip_csync2 = not service_manager.service_is_active(CSYNC2_SERVICE, cluster_node)
                 if _context.skip_csync2:
                     service_manager.stop_service(CSYNC2_SERVICE, disable=True)
@@ -2557,14 +2630,6 @@
     ocfs2_inst.join_ocfs2(peer_host)
 
 
-def join_remote_auth(node, user):
-    if os.path.exists(PCMK_REMOTE_AUTH):
-        utils.rmfile(PCMK_REMOTE_AUTH)
-    pcmk_remote_dir = os.path.dirname(PCMK_REMOTE_AUTH)
-    utils.mkdirs_owned(pcmk_remote_dir, mode=0o750, gid="haclient")
-    utils.touch(PCMK_REMOTE_AUTH)
-
-
 def remove_qdevice():
     """
     Remove qdevice service and configuration from cluster

crmsh/ui_cluster.py

@@ -331,12 +331,6 @@ def do_init(self, context, *args):
         '''
         Initialize a cluster.
         '''
-        def looks_like_hostnames(lst):
-            sectionlist = bootstrap.INIT_STAGES
-            return all(not (l.startswith('-') or l in sectionlist) for l in lst)
-        if len(args) > 0:
-            if '--dry-run' in args or looks_like_hostnames(args):
-                args = ['--yes', '--nodes'] + [arg for arg in args if arg != '--dry-run']
         parser = ArgumentParser(description="""
 Initialize a cluster from scratch. This command configures
 a complete cluster, and can also add additional cluster
@@ -471,8 +465,6 @@ def looks_like_hostnames(lst):
         if stage == "vgfs":
             stage = "ocfs2"
             logger.warning("vgfs stage was deprecated and is an alias of ocfs2 stage now")
-        if stage not in bootstrap.INIT_STAGES and stage != "":
-            parser.error("Invalid stage (%s)" % (stage))
 
         if options.qnetd_addr_input:
             if not ServiceManager().service_is_available("corosync-qdevice.service"):
@@ -547,12 +539,11 @@ def do_join(self, context, *args):
         stage = ""
         if len(args) == 1:
             stage = args[0]
-        if stage not in ("ssh", "csync2", "ssh_merge", "cluster", ""):
-            parser.error("Invalid stage (%s)" % (stage))
 
         join_context = bootstrap.Context.set_context(options)
         join_context.ui_context = context
         join_context.stage = stage
+        join_context.cluster_is_running = ServiceManager(sh.ClusterShellAdaptorForLocalShell(sh.LocalShell())).service_is_active("pacemaker.service")
         join_context.type = "join"
         join_context.validate_option()
 

crmsh/utils.py

@@ -468,14 +468,6 @@ def chmod(path, mod):
             fatal("Failed to chmod {}: {}".format(path, err))
 
 
-def touch(file_name):
-    rc, out, err = ShellUtils().get_stdout_stderr("touch " + file_name, no_reg=True)
-    if rc != 0:
-        rc, out, err = ShellUtils().get_stdout_stderr("sudo touch " + file_name, no_reg=True)
-        if rc != 0:
-            fatal("Failed create file {}: {}".format(file_name, err))
-
-
 def copy_local_file(src, dest):
     try:
         shutil.copyfile(src, dest)

test/features/bootstrap_bugs.feature

@@ -4,6 +4,31 @@ Feature: Regression test for bootstrap bugs
   Tag @clean means need to stop cluster service if the service is available
   Need nodes: hanode1 hanode2 hanode3
 
+  @clean
+  Scenario: Stages dependency (bsc#1175865)
+    Given   Cluster service is "stopped" on "hanode1"
+    And     Cluster service is "stopped" on "hanode2"
+    When    Try "crm cluster init cluster -y" on "hanode1"
+    Then    Except "ERROR: cluster.init: Please run 'ssh' stage first"
+    When    Run "crm cluster init ssh -y" on "hanode1"
+    When    Try "crm cluster init cluster -y" on "hanode1"
+    Then    Except "ERROR: cluster.init: Please run 'csync2' stage first"
+    When    Run "crm cluster init csync2 -y" on "hanode1"
+    When    Try "crm cluster init cluster -y" on "hanode1"
+    Then    Except "ERROR: cluster.init: Please run 'corosync' stage first"
+    When    Run "crm cluster init corosync -y" on "hanode1"
+    When    Run "crm cluster init cluster -y" on "hanode1"
+    Then    Cluster service is "started" on "hanode1"
+
+    When    Try "crm cluster join cluster -c hanode1 -y" on "hanode2"
+    Then    Except "ERROR: cluster.join: Please run 'ssh' stage first"
+    When    Try "crm cluster join ssh -c hanode1 -y" on "hanode2"
+    When    Try "crm cluster join cluster -c hanode1 -y" on "hanode2"
+    Then    Except "ERROR: cluster.join: Please run 'csync2' stage first"
+    When    Try "crm cluster join csync2 -c hanode1 -y" on "hanode2"
+    When    Try "crm cluster join cluster -c hanode1 -y" on "hanode2"
+    Then    Cluster service is "started" on "hanode2"
+
   @clean
   Scenario: Set placement-strategy value as "default"(bsc#1129462)
     Given   Cluster service is "stopped" on "hanode1"

test/features/bootstrap_options.feature

@@ -41,6 +41,15 @@ Feature: crmsh bootstrap process - options
     When    Try "crm cluster init sbd -N hanode1 -N hanode2 -y" on "hanode1"
     Then    Expected "Can't use -N/--nodes option and stage(sbd) together" in stderr
 
+  @clean
+  Scenario: Stage validation
+    When    Try "crm cluster init fdsf -y" on "hanode1"
+    Then    Expected "Invalid stage: fdsf(available stages: ssh, csync2, corosync, sbd, cluster, ocfs2, admin, qdevice)" in stderr
+    When    Try "crm cluster join fdsf -y" on "hanode1"
+    Then    Expected "Invalid stage: fdsf(available stages: ssh, csync2, ssh_merge, cluster)" in stderr
+    When    Try "crm cluster join ssh -y" on "hanode1"
+    Then    Expected "Can't use stage(ssh) without specifying cluster node" in stderr
+
   @clean
   Scenario: Init whole cluster service on node "hanode1" using "--node" option
     Given   Cluster service is "stopped" on "hanode1"
@@ -51,6 +60,9 @@ Feature: crmsh bootstrap process - options
     And     Online nodes are "hanode1 hanode2"
     And     Show cluster status on "hanode1"
 
+    When    Try "crm cluster init cluster -y" on "hanode1"
+    Then    Expected "Cluster is active, can't run 'cluster' stage" in stderr
+
   @clean
   Scenario: Bind specific network interface using "-i" option
     Given   Cluster service is "stopped" on "hanode1"
@@ -96,6 +108,9 @@ Feature: crmsh bootstrap process - options
     And     Cluster virtual IP is "@vip.0"
     And     Show cluster status on "hanode1"
 
+    When    Try "crm cluster init cluster -y" on "hanode1"
+    Then    Expected "Cluster is active, can't run 'cluster' stage" in stderr
+
   @clean
   Scenario: Init cluster service with udpu using "-u" option
     Given   Cluster service is "stopped" on "hanode1"

test/features/qdevice_validate.feature

@@ -110,7 +110,7 @@ Feature: corosync qdevice/qnetd options validate
   Scenario: Run qdevice stage on inactive cluster node
     Given   Cluster service is "stopped" on "hanode1"
     When    Try "crm cluster init qdevice --qnetd-hostname=qnetd-node"
-    Then    Except "ERROR: cluster.init: Cluster is inactive - can't run qdevice stage"
+    Then    Except "ERROR: cluster.init: Cluster is inactive, can't run 'qdevice' stage"
 
   @clean
   Scenario: Run qdevice stage but miss "--qnetd-hostname" option

test/unittests/test_bootstrap.py

@@ -564,7 +564,7 @@ def test_join_ssh(
             mock_get_node_cononical_hostname,
             mock_detect_cluster_service_on_node
     ):
-        bootstrap._context = mock.Mock(current_user="bob", default_nic="eth1", use_ssh_agent=False)
+        bootstrap._context = mock.Mock(current_user="bob", default_nic="eth1", use_ssh_agent=False, stage=None)
         mock_swap.return_value = None
         mock_ssh_copy_id.return_value = 0
         mock_get_node_cononical_hostname.return_value='node1'