[ARVADOS] created: 1.1.4-347-g32568bb
Git user
git at public.curoverse.com
Fri Jun 1 14:54:42 EDT 2018
at 32568bbde8ad549452ccb57f10bb54672fda6fa6 (commit)
commit 32568bbde8ad549452ccb57f10bb54672fda6fa6
Author: Tom Clegg <tclegg at veritasgenetics.com>
Date: Fri Jun 1 13:35:10 2018 -0400
12061: Disable nodemanager integration tests.
Arvados-DCO-1.1-Signed-off-by: Tom Clegg <tclegg at veritasgenetics.com>
diff --git a/build/run-tests.sh b/build/run-tests.sh
index 8a8f5b6..2b280c7 100755
--- a/build/run-tests.sh
+++ b/build/run-tests.sh
@@ -270,6 +270,8 @@ declare -a failures
declare -A skip
declare -A testargs
skip[apps/workbench_profile]=1
+# nodemanager_integration tests are not reliable, see #12061.
+skip[services/nodemanager_integration]=1
while [[ -n "$1" ]]
do
commit 6c5fc89c06338ea829f2e4f2a7e0529548890efd
Author: Tom Clegg <tclegg at veritasgenetics.com>
Date: Fri Jun 1 13:36:43 2018 -0400
12061: Remove unreliable check.
"Shutdown because state=down" is possible during a race, when
nodemanager logs "Shutdown success" and polls sinfo again before the
test suite has finished updating the sinfo stub.
Arvados-DCO-1.1-Signed-off-by: Tom Clegg <tclegg at veritasgenetics.com>
diff --git a/services/nodemanager/tests/integration_test.py b/services/nodemanager/tests/integration_test.py
index 3f616c6..a8429e1 100755
--- a/services/nodemanager/tests/integration_test.py
+++ b/services/nodemanager/tests/integration_test.py
@@ -113,7 +113,8 @@ def node_busy(g):
def node_shutdown(g):
global compute_nodes
- del compute_nodes[g.group(1)]
+ if g.group(1) in compute_nodes:
+ del compute_nodes[g.group(1)]
return 0
def jobs_req(g):
@@ -321,7 +322,6 @@ def main():
],
# Checks (things that shouldn't happen)
{
- r".*Suggesting shutdown because node state is \('down', .*\)": fail,
r".*Cloud node (\S+) is now paired with Arvados node (\S+) with hostname (\S+)": partial(expect_count, 4),
r".*Setting node quota.*": fail,
},
@@ -347,7 +347,6 @@ def main():
],
# Checks (things that shouldn't happen)
{
- r".*Suggesting shutdown because node state is \('down', .*\)": fail,
r".*Cloud node (\S+) is now paired with Arvados node (\S+) with hostname (\S+)": partial(expect_count, 2),
r".*Sending create_node request.*": partial(expect_count, 5)
},
@@ -384,7 +383,6 @@ def main():
],
# Checks (things that shouldn't happen)
{
- r".*Suggesting shutdown because node state is \('down', .*\)": fail,
r".*Cloud node (\S+) is now paired with Arvados node (\S+) with hostname (\S+)": partial(expect_count, 6),
r".*Sending create_node request.*": partial(expect_count, 9)
},
@@ -444,7 +442,6 @@ def main():
],
# Checks (things that shouldn't happen)
{
- r".*Suggesting shutdown because node state is \('down', .*\)": fail,
r".*Cloud node (\S+) is now paired with Arvados node (\S+) with hostname (\S+)": partial(expect_count, 1),
r".*Setting node quota.*": fail,
},
@@ -465,7 +462,6 @@ def main():
],
# Checks (things that shouldn't happen)
{
- r".*Suggesting shutdown because node state is \('down', .*\)": fail,
r".*Cloud node (\S+) is now paired with Arvados node (\S+) with hostname (\S+)": partial(expect_count, 1),
r".*Setting node quota.*": fail,
},
commit 3108e6a627f4413819af8d2eb006016f1695b9a8
Author: Tom Clegg <tclegg at veritasgenetics.com>
Date: Fri Jun 1 12:50:02 2018 -0400
12061: Reduce per-test timeout.
Arvados-DCO-1.1-Signed-off-by: Tom Clegg <tclegg at veritasgenetics.com>
diff --git a/services/nodemanager/tests/integration_test.py b/services/nodemanager/tests/integration_test.py
index 2de3769..3f616c6 100755
--- a/services/nodemanager/tests/integration_test.py
+++ b/services/nodemanager/tests/integration_test.py
@@ -175,8 +175,8 @@ def run_test(name, actions, checks, driver_class, jobs, provider):
driver_class=driver_class,
ssh_key=os.path.join(fake_slurm, "id_rsa.pub")))
- # Tests must complete in less than 3 minutes.
- timeout = time.time() + 180
+ # Tests must complete in less than 30 seconds.
+ timeout = time.time() + 30
terminated = False
# Now start node manager
commit e1ee43ee90b7f6e6189930cb031448014d0cf638
Author: Tom Clegg <tclegg at veritasgenetics.com>
Date: Fri Jun 1 12:49:42 2018 -0400
12061: Fix nodes marked busy while expecting shutdown decision.
Arvados-DCO-1.1-Signed-off-by: Tom Clegg <tclegg at veritasgenetics.com>
diff --git a/services/nodemanager/tests/integration_test.py b/services/nodemanager/tests/integration_test.py
index 4aa8c9e..2de3769 100755
--- a/services/nodemanager/tests/integration_test.py
+++ b/services/nodemanager/tests/integration_test.py
@@ -106,18 +106,6 @@ def node_paired(g):
return 0
-def remaining_jobs(g):
- update_script(os.path.join(fake_slurm, "sinfo"), "#!/bin/sh\n" +
- "\n".join("echo '%s|alloc|(null)'" % (v) for k,v in compute_nodes.items()))
-
- for k,v in all_jobs.items():
- all_jobs[k] = "Running"
-
- set_squeue(g)
-
- return 0
-
-
def node_busy(g):
update_script(os.path.join(fake_slurm, "sinfo"), "#!/bin/sh\n" +
"\n".join("echo '%s|idle|(null)'" % (v) for k,v in compute_nodes.items()))
@@ -353,7 +341,7 @@ def main():
(r".*Cloud node (\S+) is now paired with Arvados node (\S+) with hostname (\S+)", node_paired),
(r".*Cloud node (\S+) is now paired with Arvados node (\S+) with hostname (\S+)", node_paired),
(r".*ComputeNodeMonitorActor\..*\.([^[]*).*Not eligible for shut down because node state is \('busy', 'open', .*\)", node_busy),
- (r".*ComputeNodeMonitorActor\..*\.([^[]*).*Suggesting shutdown because node state is \('idle', 'open', .*\)", remaining_jobs),
+ (r".*ComputeNodeMonitorActor\..*\.([^[]*).*Suggesting shutdown because node state is \('idle', 'open', .*\)", noop),
(r".*ComputeNodeShutdownActor\..*\.([^[]*).*Shutdown success", node_shutdown),
(r".*ComputeNodeShutdownActor\..*\.([^[]*).*Shutdown success", node_shutdown)
],
@@ -379,7 +367,7 @@ def main():
(r".*Cloud node (\S+) is now paired with Arvados node (\S+) with hostname (\S+)", node_paired),
(r".*Cloud node (\S+) is now paired with Arvados node (\S+) with hostname (\S+)", node_paired),
(r".*ComputeNodeMonitorActor\..*\.([^[]*).*Not eligible for shut down because node state is \('busy', 'open', .*\)", node_busy),
- (r".*ComputeNodeMonitorActor\..*\.([^[]*).*Suggesting shutdown because node state is \('idle', 'open', .*\)", remaining_jobs),
+ (r".*ComputeNodeMonitorActor\..*\.([^[]*).*Suggesting shutdown because node state is \('idle', 'open', .*\)", noop),
(r".*ComputeNodeShutdownActor\..*\.([^[]*).*Shutdown success", node_shutdown),
(r".*ComputeNodeShutdownActor\..*\.([^[]*).*Shutdown success", node_shutdown),
(r".*sending request", jobs_req),
-----------------------------------------------------------------------
hooks/post-receive
--
More information about the arvados-commits
mailing list