include correct soak updates for 2.6

Required-githooks: true

Signed-off-by: Maureen Jean <[email protected]>

src/tests/ftest/soak/harassers.yaml

@@ -3,7 +3,7 @@ hosts:
   test_servers: 8
   # servers if a server partition is defined
   # server_partition: daos_server
-  client_partition: daos_client
+  # client_partition: daos_client
   # client_reservation: daos-test
 orterun:
   allow_run_as_root: true

src/tests/ftest/soak/smoke.yaml

@@ -3,7 +3,7 @@ hosts:
   test_servers: 4
   # servers if a server partition is defined
   # server_partition: daos_server
-  client_partition: daos_client
+  # client_partition: daos_client
   # client_reservation: daos-test
 orterun:
   allow_run_as_root: true

src/tests/ftest/soak/stress.yaml

@@ -3,7 +3,7 @@ hosts:
   test_servers: 8
   # servers if a server partition is defined
   # server_partition: daos_server
-  client_partition: daos_client
+  # client_partition: daos_client
   # client_reservation: daos-test
 orterun:
   allow_run_as_root: true
@@ -138,7 +138,6 @@ ior_stress:
     mount_dir: "/tmp/soak_dfuse_ior/"
     disable_caching: true
     thread_count: 8
-    cores: '0-7'
 fio_stress:
   api:
     - POSIX
@@ -180,7 +179,6 @@ fio_stress:
     mount_dir: "/tmp/soak_dfuse_fio/"
     disable_caching: true
     thread_count: 8
-    cores: '0-7'
 daos_racer:
   runtime: 120
 vpic_stress:
@@ -217,7 +215,6 @@ lammps_stress:
     mount_dir: "/tmp/soak_dfuse_lammps/"
     disable_caching: true
     thread_count: 8
-    cores: '0-7'
   oclass:
     - ["EC_2P1GX", "RP_2GX"]
 mdtest_stress:
@@ -258,7 +255,6 @@ mdtest_stress:
     mount_dir: "/tmp/soak_dfuse_mdtest/"
     disable_caching: true
     thread_count: 8
-    cores: '0-7'
 macsio_stress:
   job_timeout: 30
   nodesperjob:
@@ -289,7 +285,6 @@ macsio_stress:
     mount_dir: "/tmp/soak_dfuse_macsio/"
     disable_caching: true
     thread_count: 8
-    cores: '0-7'
 datamover_stress:
   job_timeout: 10
   nodesperjob:

src/tests/ftest/util/soak_test_base.py

@@ -145,6 +145,9 @@ def pre_tear_down(self):
             if not run_local(self.log, cmd, timeout=120).passed:
                 # Exception was raised due to a non-zero exit status
                 errors.append(f"Failed to cancel jobs {self.failed_job_id_list}")
+        elif self.job_scheduler != "slurm":
+            cmd = "pkill jobscript"
+            run_remote(self.log, self.hostlist_clients, cmd)
         if self.all_failed_jobs:
             errors.append("SOAK FAILED: The following jobs failed {} ".format(
                 " ,".join(str(j_id) for j_id in self.all_failed_jobs)))

src/tests/ftest/util/soak_utils.py

@@ -22,7 +22,7 @@
 from daos_racer_utils import DaosRacerCommand
 from data_mover_utils import DcpCommand, FsCopy
 from dfuse_utils import get_dfuse
-from dmg_utils import get_storage_query_device_info
+from dmg_utils import get_storage_query_device_info, get_storage_query_device_uuids
 from duns_utils import format_path
 from exception_utils import CommandFailure
 from fio_utils import FioCommand
@@ -573,8 +573,10 @@ def launch_vmd_identify_check(self, name, results, args):
         results (queue): multiprocessing queue
         args (queue): multiprocessing queue
     """
+    # pylint: disable=too-many-nested-blocks
     status = True
     failing_vmd = []
+    dmg = self.get_dmg_command().copy()
     device_info = get_storage_query_device_info(self.dmg_command)
     uuid_list = [device['uuid'] for device in device_info]
     # limit the number of leds to blink to 1024
@@ -583,22 +585,27 @@ def launch_vmd_identify_check(self, name, results, args):
     else:
         uuids = uuid_list
     self.log.info("VMD device UUIDs: %s", uuids)
-
-    for uuid in uuids:
-        # Blink led
-        self.dmg_command.storage_led_identify(ids=uuid, timeout=2)
-        # check if led is blinking
-        result = self.dmg_command.storage_led_check(ids=uuid)
-        # determine if leds are blinking as expected
-        for value in list(result['response']['host_storage_map'].values()):
-            if value['storage']['smd_info']['devices']:
-                for device in value['storage']['smd_info']['devices']:
-                    if device['ctrlr']['led_state'] != "QUICK_BLINK":
-                        failing_vmd.append([device['ctrlr']['pci_addr'], value['hosts']])
-                        status = False
-    # reset leds to previous state
-    for uuid in uuids:
-        self.dmg_command.storage_led_identify(ids=uuid, reset=True)
+    host_uuids = get_storage_query_device_uuids(self.dmg_command)
+    for host, uuid_dict in host_uuids.items():
+        uuid_list = sorted(uuid_dict.keys())
+        self.log.info("Devices on host %s: %s", host, uuid_list)
+        # Now check whether the random uuid belongs to a particular host.
+        for uuid in uuids:
+            if uuid in uuid_list:
+                dmg.hostlist = host
+                # Blink led
+                dmg.storage_led_identify(ids=uuid, timeout=2)
+                # check if led is blinking
+                result = dmg.storage_led_check(ids=uuid)
+                # determine if leds are blinking as expected
+                for value in list(result['response']['host_storage_map'].values()):
+                    if value['storage']['smd_info']['devices']:
+                        for device in value['storage']['smd_info']['devices']:
+                            if device['ctrlr']['led_state'] != "QUICK_BLINK":
+                                failing_vmd.append([device['ctrlr']['pci_addr'], value['hosts']])
+                                status = False
+                            # reset leds to previous state
+                            dmg.storage_led_identify(ids=uuid, reset=True)
 
     params = {"name": name,
               "status": status,