cephadm: avoid crashing on expected non-zero exit

- Avoid crashing when a call out to an external program expectedly does
  not return exit status zero.

There are programs that communicate other information than error/no
error through exit status. E.g. `systemctl status` will return different
exit codes depending on the actual status of the units in question.
In cases where this is expected crashing with a RuntimeError exception
is inappropriate and should be avoided.

Fixes: https://tracker.ceph.com/issues/55117
Signed-off-by: Moritz Röhrich <moritz.rohrich@suse.com>

diff --git a/qa/workunits/cephadm/test_cephadm.sh b/qa/workunits/cephadm/test_cephadm.sh
index 643065b73b1c79b2a8d213cd81558ab24b34433c..31ebf0a5dd98d1791b8c3fb56d86bacf9c41f7a2 100755 (executable)
--- a/qa/workunits/cephadm/test_cephadm.sh
+++ b/qa/workunits/cephadm/test_cephadm.sh
@@ -77,6 +77,22 @@ function expect_false()
         if eval "$@"; then return 1; else return 0; fi
 }
 
+# expect_return_code $expected_code $command ...
+function expect_return_code()
+{
+  set -x
+  local expected_code="$1"
+  shift
+  local command="$@"
+
+  set +e
+  eval "$command"
+  local return_code="$?"
+  set -e
+
+  if [ ! "$return_code" -eq "$expected_code" ]; then return 1; else return 0; fi
+}
+
 function is_available()
 {
     local name="$1"
@@ -370,6 +386,10 @@ $CEPHADM unit --fsid $FSID --name mon.a -- disable
 expect_false $CEPHADM unit --fsid $FSID --name mon.a -- is-enabled
 $CEPHADM unit --fsid $FSID --name mon.a -- enable
 $CEPHADM unit --fsid $FSID --name mon.a -- is-enabled
+$CEPHADM unit --fsid $FSID --name mon.a -- status
+$CEPHADM unit --fsid $FSID --name mon.a -- stop
+expect_return_code 3 $CEPHADM unit --fsid $FSID --name mon.a -- status
+$CEPHADM unit --fsid $FSID --name mon.a -- start
 
 ## shell
 $CEPHADM shell --fsid $FSID -- true

diff --git a/src/cephadm/cephadm b/src/cephadm/cephadm
index 06720c0caa48a10de3067bc6e8a5d0573d6644e6..c40d2165f26b4f5a391f306378fb1434f55b600a 100755 (executable)
--- a/src/cephadm/cephadm
+++ b/src/cephadm/cephadm
@@ -5674,19 +5674,19 @@ def command_ceph_volume(ctx):
 
 @infer_fsid
 def command_unit(ctx):
-    # type: (CephadmContext) -> None
+    # type: (CephadmContext) -> int
     if not ctx.fsid:
         raise Error('must pass --fsid to specify cluster')
 
     unit_name = get_unit_name_by_daemon_name(ctx, ctx.fsid, ctx.name)
 
-    call_throws(ctx, [
-        'systemctl',
-        ctx.command,
-        unit_name],
+    _, _, code = call(
+        ctx,
+        ['systemctl', ctx.command, unit_name],
         verbosity=CallVerbosity.VERBOSE,
         desc=''
     )
+    return code
 
 ##################################