Infinite mea culpa for the number of problems with this test.
This change bumps the acceptable RSS increase threshold from 3% to 50% when spamming the test node with junk 4MB messages. On @MarcoFalke's macOS test build we see RSS grow ~14% from ~71MB to 81MB, so a 50% increase threshold should be more than sufficient to avoid spurious failures.
65 | @@ -66,7 +66,7 @@ def run_test(self): 66 | msg_at_size = msg_unrecognized("b" * valid_data_limit) 67 | assert len(msg_at_size.serialize()) == msg_limit 68 | 69 | - with node.assert_memory_usage_stable(perc_increase_allowed=0.03): 70 | + with node.assert_memory_usage_stable(perc_increase_allowed=0.5):
Doesn't percent mean it is denominated in hundreds? So maybe rename it at the same time to increase_allowed? And maybe require it to be a named arg with , *, increase_allowed in the signature?
Good suggestions; added in a subsequent commit.
Post-merge (that was quick btw, but I guess ok since it's only tests): I assume error is unrelated?
2018-11-27T07:29:34.685000Z TestFramework (INFO): Initializing test directory /var/folders/b8/znqr8hj918772gfmd875gzgdd3ypz1/T/testzl4oubll
2018-11-27T07:29:35.945000Z TestFramework (INFO): Sending a bunch of large, junk messages to test memory exhaustion. May take a bit...
2018-11-27T07:29:40.680000Z TestFramework (INFO): Waiting for node to drop junk messages.
Fatal write error on socket transport
protocol: <test_framework.mininode.P2PDataStore object at 0x10f38d240>
transport: <_SelectorSocketTransport fd=10 read=polling write=<idle, bufsize=0>>
Traceback (most recent call last):
File "/usr/local/Cellar/python/3.7.1/Frameworks/Python.framework/Versions/3.7/lib/python3.7/asyncio/selector_events.py", line 868, in write
n = self._sock.send(data)
OSError: [Errno 41] Protocol wrong type for socket
2018-11-27T07:29:40.799000Z TestFramework.mininode (WARNING): Connection lost to 127.0.0.1:14387 due to [Errno 41] Protocol wrong type for socket
2018-11-27T07:29:40.899000Z TestFramework (INFO): Sending a message with incorrect size of 2
2018-11-27T07:29:40.956000Z TestFramework.mininode (WARNING): Connection lost to 127.0.0.1:14387 due to [Errno 54] Connection reset by peer
2018-11-27T07:29:41.061000Z TestFramework (INFO): Sending a message with incorrect size of 77
2018-11-27T07:29:41.222000Z TestFramework (INFO): Sending a message with incorrect size of 78
2018-11-27T07:29:41.384000Z TestFramework (INFO): Sending a message with incorrect size of 79
2018-11-27T07:29:41.761000Z TestFramework (INFO): Stopping nodes
2018-11-27T07:29:42.346000Z TestFramework (INFO): Cleaning up /var/folders/b8/znqr8hj918772gfmd875gzgdd3ypz1/T/testzl4oubll on exit
2018-11-27T07:29:42.346000Z TestFramework (INFO): Tests successful
I met the error that @kallewoof pointed out above, too. (macbook pro/macos 10.13.6)
p2p_invalid_messages.py says Tests successful, but test_runner.py judges that it Failed.
$ ./test/functional/test_runner.py p2p_invalid_messages.py
Temporary test directory at /var/folders/gy/zxy8j2k17yxfpdbs_td03mwjzvzqs2/T/test_runner_₿_🏃_20181127_164919
1/1 - p2p_invalid_messages.py failed, Duration: 6 s
stdout:
2018-11-27T07:49:20.102000Z TestFramework (INFO): Initializing test directory /var/folders/gy/zxy8j2k17yxfpdbs_td03mwjzvzqs2/T/test_runner_₿_🏃_20181127_164919/p2p_invalid_messages_0
2018-11-27T07:49:21.079000Z TestFramework (INFO): Sending a bunch of large, junk messages to test memory exhaustion. May take a bit...
2018-11-27T07:49:24.885000Z TestFramework (INFO): Waiting for node to drop junk messages.
2018-11-27T07:49:24.989000Z TestFramework.mininode (WARNING): Connection lost to 127.0.0.1:11000 due to [Errno 41] Protocol wrong type for socket
2018-11-27T07:49:25.098000Z TestFramework (INFO): Sending a message with incorrect size of 2
2018-11-27T07:49:25.153000Z TestFramework.mininode (WARNING): Connection lost to 127.0.0.1:11000 due to [Errno 54] Connection reset by peer
2018-11-27T07:49:25.258000Z TestFramework (INFO): Sending a message with incorrect size of 77
2018-11-27T07:49:25.314000Z TestFramework.mininode (WARNING): Connection lost to 127.0.0.1:11000 due to [Errno 54] Connection reset by peer
2018-11-27T07:49:25.420000Z TestFramework (INFO): Sending a message with incorrect size of 78
2018-11-27T07:49:25.579000Z TestFramework (INFO): Sending a message with incorrect size of 79
2018-11-27T07:49:25.960000Z TestFramework (INFO): Stopping nodes
2018-11-27T07:49:26.446000Z TestFramework (INFO): Cleaning up /var/folders/gy/zxy8j2k17yxfpdbs_td03mwjzvzqs2/T/test_runner_₿_🏃_20181127_164919/p2p_invalid_messages_0 on exit
2018-11-27T07:49:26.446000Z TestFramework (INFO): Tests successful
stderr:
Fatal write error on socket transport
protocol: <test_framework.mininode.P2PDataStore object at 0x10ae10a58>
transport: <_SelectorSocketTransport fd=10 read=polling write=<idle, bufsize=0>>
Traceback (most recent call last):
File "/usr/local/Cellar/python/3.7.1/Frameworks/Python.framework/Versions/3.7/lib/python3.7/asyncio/selector_events.py", line 868, in write
n = self._sock.send(data)
OSError: [Errno 41] Protocol wrong type for socket
TEST | STATUS | DURATION
p2p_invalid_messages.py | ✖ Failed | 6 s
ALL | ✖ Failed | 6 s (accumulated)
Runtime: 6 s
Is that a deterministic failure due to one of the commits in this pull request? I doubt it, since I have seen it before, but please open a new issue about this.
According to some other reports asyncio has issues with large buffers on macOS...
Yeah this could be an error that would happen if the original error did not (i.e. the situation post-merge).