From: Sebastian Wagner <sebastian.wagner@suse.com>
Date: Wed, 26 Aug 2020 12:53:09 +0000 (+0200)
Subject: mgr/cephadm: Call cephadm with --container-image
X-Git-Tag: v15.2.9~122^2~44^2~37
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=a05839c2294188dc7a66d764f6ed286d540995f5;p=ceph.git

mgr/cephadm: Call cephadm with --container-image

The kernel treats any process with PID 1 different. Especially
it does not generate a core dump. Call podman / docker with
--init in order to get core dumps.

In addition, we can now properly reap zombies processes.

Fixes: https://tracker.ceph.com/issues/44231

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>
(cherry picked from commit 533931505272d215a1b4fa40b3bf235cf2110102)
---

diff --git a/doc/man/8/cephadm.rst b/doc/man/8/cephadm.rst
index 982a01f1bfd..5a1da7a040e 100644
--- a/doc/man/8/cephadm.rst
+++ b/doc/man/8/cephadm.rst
@@ -28,6 +28,7 @@ Synopsis
 | **cephadm** **adopt** [-h] --name NAME --style STYLE [--cluster CLUSTER]
 |                       [--legacy-dir LEGACY_DIR] [--config-json CONFIG_JSON]
 |                       [--skip-firewalld] [--skip-pull]
+|                       [--container-init]
 
 | **cephadm** **rm-daemon** [-h] --name NAME --fsid FSID [--force]
 |                           [--force-delete-data]
diff --git a/src/cephadm/cephadm b/src/cephadm/cephadm
index 1b524217cba..12c9ff99cbf 100755
--- a/src/cephadm/cephadm
+++ b/src/cephadm/cephadm
@@ -3028,6 +3028,9 @@ def command_bootstrap():
         cli(['config', 'set', 'mgr', 'mgr/cephadm/registry_username', args.registry_username, '--force'])
         cli(['config', 'set', 'mgr', 'mgr/cephadm/registry_password', args.registry_password, '--force'])
 
+    if args.container_init:
+        cli(['config', 'set', 'mgr', 'mgr/cephadm/container_init', args.container_init, '--force'])
+
     if not args.skip_dashboard:
         # Configure SSL port (cephadm only allows to configure dashboard SSL port)
         # if the user does not want to use SSL he can change this setting once the cluster is up
@@ -5384,6 +5387,10 @@ def _get_parser():
         '--force-start',
         action='store_true',
         help="start newly adoped daemon, even if it wasn't running previously")
+    parser_adopt.add_argument(
+        '--container-init',
+        action='store_true',
+        help='Run podman/docker with `--init`')
 
     parser_rm_daemon = subparsers.add_parser(
         'rm-daemon', help='remove daemon instance')
diff --git a/src/pybind/mgr/cephadm/module.py b/src/pybind/mgr/cephadm/module.py
index 9471ac84a7f..eb3fe837008 100644
--- a/src/pybind/mgr/cephadm/module.py
+++ b/src/pybind/mgr/cephadm/module.py
@@ -214,6 +214,12 @@ class CephadmOrchestrator(orchestrator.Orchestrator, MgrModule):
                          'can allow debugging daemons that encounter problems '
                          'at runtime.',
         },
+        {
+            'name': 'container_init',
+            'type': 'bool',
+            'default': False,
+            'desc': 'Run podman/docker with `--init`',
+        },
         {
             'name': 'prometheus_alerts_path',
             'type': 'str',
@@ -295,6 +301,7 @@ class CephadmOrchestrator(orchestrator.Orchestrator, MgrModule):
             self.warn_on_stray_daemons = True
             self.warn_on_failed_host_check = True
             self.allow_ptrace = False
+            self.container_init = False
             self.prometheus_alerts_path = ''
             self.migration_current = None
             self.config_dashboard = True
@@ -1168,6 +1175,10 @@ To check that the host is reachable:
 
             if not no_fsid:
                 final_args += ['--fsid', self._cluster_fsid]
+
+            if self.container_init:
+                final_args += ['--container-init']
+
             final_args += args
 
             self.log.debug('args: %s' % (' '.join(final_args)))