qa/tasks/mgr: clean crash reports before waiting for clean

otherwise we have following warning in health report

{"status":"HEALTH_WARN","checks":{"RECENT_MGR_MODULE_CRASH":{"severity":"HEALTH_WARN","summary":{"message":"1 mgr modules have recently crashed","count":1},"muted":false}},"mutes":[]}

and it does not disappear after the test waits for 30 seconds.
and the tasks.mgr.test_module_selftest.TestModuleSelftest test
fails like:

2021-07-21T09:59:52.560 INFO:tasks.cephfs_test_runner:======================================================================
2021-07-21T09:59:52.561 INFO:tasks.cephfs_test_runner:ERROR: test_module_commands (tasks.mgr.test_module_selftest.TestModuleSelftest)
2021-07-21T09:59:52.561 INFO:tasks.cephfs_test_runner:----------------------------------------------------------------------
2021-07-21T09:59:52.561 INFO:tasks.cephfs_test_runner:Traceback (most recent call last):
2021-07-21T09:59:52.562 INFO:tasks.cephfs_test_runner:  File "/home/teuthworker/src/git.ceph.com_ceph-c_6a5d5abc027f706687dec92f92ff6fc6f074d2ae/qa/tasks/mgr/test_module_selftest.py", line 201, in
test_mo
dule_commands
2021-07-21T09:59:52.562 INFO:tasks.cephfs_test_runner:    self.wait_for_health_clear(timeout=30)
2021-07-21T09:59:52.562 INFO:tasks.cephfs_test_runner:  File "/home/teuthworker/src/git.ceph.com_ceph-c_6a5d5abc027f706687dec92f92ff6fc6f074d2ae/qa/tasks/ceph_test_case.py", line 172, in
wait_for_health_c
lear
2021-07-21T09:59:52.563 INFO:tasks.cephfs_test_runner:    self.wait_until_true(is_clear, timeout)
2021-07-21T09:59:52.563 INFO:tasks.cephfs_test_runner:  File "/home/teuthworker/src/git.ceph.com_ceph-c_6a5d5abc027f706687dec92f92ff6fc6f074d2ae/qa/tasks/ceph_test_case.py", line 209, in
wait_until_true
2021-07-21T09:59:52.563 INFO:tasks.cephfs_test_runner:    raise TestTimeoutError("Timed out after {0}s and {1} retries".format(elapsed, retry_count))
2021-07-21T09:59:52.564 INFO:tasks.cephfs_test_runner:tasks.ceph_test_case.TestTimeoutError: Timed out after 30s and 0 retries

in this change, the crash reports are nuked right after
we see the warning, so that we can have a clean health
report.

Fixes: https://tracker.ceph.com/issues/51743
Signed-off-by: Kefu Chai <kchai@redhat.com>

diff --git a/qa/tasks/mgr/test_module_selftest.py b/qa/tasks/mgr/test_module_selftest.py
index 523a59f38bd9272e771df9e8d35c12acf0a9a6f1..0f64adbd88587b118059cffe3882fac557fb8842 100644 (file)
--- a/qa/tasks/mgr/test_module_selftest.py
+++ b/qa/tasks/mgr/test_module_selftest.py
@@ -194,7 +194,10 @@ class TestModuleSelftest(MgrTestCase):
         self.wait_for_health(
             "Module 'selftest' has failed: Synthetic exception in serve",
             timeout=30)
-
+        # prune the crash reports, so that the health report is back to
+        # clean
+        self.mgr_cluster.mon_manager.raw_cluster_cmd(
+            "crash", "prune", "0")
         self.mgr_cluster.mon_manager.raw_cluster_cmd(
             "mgr", "module", "disable", "selftest")