[SPARK-25253][PYSPARK] Refactor local connection & auth code #22247
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
Author
|
Test build #95303 has finished for PR 22247 at commit
|
HyukjinKwon
reviewed
Aug 28, 2018
python/pyspark/java_gateway.py
Outdated
| raise Exception("Unexpected reply from iterator server.") | ||
|
|
||
|
|
||
| def local_connect_and_auth(sock_info): |
Member
There was a problem hiding this comment.
@squito, not a big deal but how about local_connect_and_auth (port, auth_secret) and ..
(sockfile, sock) = local_connect_and_auth(port, auth_secret)(sock_file, _) = local_connect_and_auth(java_port, auth_secret)port, auth_secret = sock_info
(sockfile, sock) = local_connect_and_auth(port, auth_secret)or
(sockfile, sock) = local_connect_and_auth(*sock_info)
Contributor
Author
There was a problem hiding this comment.
done, thanks -- I used the varargs version *sock_info for the last one, i forgot about that in python.
HyukjinKwon
reviewed
Aug 28, 2018
python/pyspark/java_gateway.py
Outdated
| # Support for both IPv4 and IPv6. | ||
| # On most of IPv6-ready systems, IPv6 will take precedence. | ||
| for res in socket.getaddrinfo("127.0.0.1", port, socket.AF_UNSPEC, socket.SOCK_STREAM): | ||
| af, socktype, proto, canonname, sa = res |
Member
There was a problem hiding this comment.
nit: af, socktype, proto, canonname, sa = res -> af, socktype, proto, _, sa = res
HyukjinKwon
approved these changes
Aug 28, 2018
HyukjinKwon
reviewed
Aug 28, 2018
| auth_secret = os.environ["PYTHON_WORKER_FACTORY_SECRET"] | ||
| sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM) | ||
| sock.connect(("127.0.0.1", java_port)) | ||
| sock_file = sock.makefile("rwb", 65536) |
Member
There was a problem hiding this comment.
I quickly tested and seems working fine. Please ignore this comment.
jiangxb1987
reviewed
Aug 28, 2018
| # do_server_auth() function and data serialization methods. | ||
| sockfile = sock.makefile("rwb", 65536) | ||
|
|
||
| (sockfile, sock) = local_connect_and_auth(port, auth_secret) |
Contributor
There was a problem hiding this comment.
We must set sock timeout to None to allow barrier() call blocking forever.
Contributor
Author
There was a problem hiding this comment.
good catch, thanks! updated
vanzin
reviewed
Aug 28, 2018
python/pyspark/java_gateway.py
Outdated
| sock.close() | ||
| sock = None | ||
| continue | ||
| break |
Contributor
There was a problem hiding this comment.
Slight shorter (and more "python-compliant"?):
- move the socket initialization (and the return) inside the try
- get rid of the continue
- use an else instead of the condition below
|
Test build #95354 has finished for PR 22247 at commit
|
|
Test build #95365 has finished for PR 22247 at commit
|
|
Test build #95380 has finished for PR 22247 at commit
|
Member
|
Merged to master. |
Contributor
|
@squito Thanks for the refactor! |
squito
added a commit
to squito/spark
that referenced
this pull request
Sep 12, 2018
This eliminates some duplication in the code to connect to a server on localhost to talk directly to the jvm. Also it gives consistent ipv6 and error handling. Two other incidental changes, that shouldn't matter: 1) python barrier tasks perform authentication immediately (rather than waiting for the BARRIER_FUNCTION indicator) 2) for `rdd._load_from_socket`, the timeout is only increased after authentication. Closes apache#22247 from squito/py_connection_refactor. Authored-by: Imran Rashid <[email protected]> Signed-off-by: hyukjinkwon <[email protected]> (cherry picked from commit 38391c9)
asfgit
pushed a commit
that referenced
this pull request
Sep 13, 2018
This eliminates some duplication in the code to connect to a server on localhost to talk directly to the jvm. Also it gives consistent ipv6 and error handling. Two other incidental changes, that shouldn't matter: 1) python barrier tasks perform authentication immediately (rather than waiting for the BARRIER_FUNCTION indicator) 2) for `rdd._load_from_socket`, the timeout is only increased after authentication. Closes #22247 from squito/py_connection_refactor. Authored-by: Imran Rashid <[email protected]> Signed-off-by: hyukjinkwon <[email protected]> (cherry picked from commit 38391c9)
asfgit
pushed a commit
that referenced
this pull request
Sep 25, 2018
This eliminates some duplication in the code to connect to a server on localhost to talk directly to the jvm. Also it gives consistent ipv6 and error handling. Two other incidental changes, that shouldn't matter: 1) python barrier tasks perform authentication immediately (rather than waiting for the BARRIER_FUNCTION indicator) 2) for `rdd._load_from_socket`, the timeout is only increased after authentication. Closes #22247 from squito/py_connection_refactor. Authored-by: Imran Rashid <[email protected]> Signed-off-by: hyukjinkwon <[email protected]> (cherry picked from commit 38391c9) (cherry picked from commit a2a54a5)
Willymontaz
pushed a commit
to criteo-forks/spark
that referenced
this pull request
Sep 26, 2019
This eliminates some duplication in the code to connect to a server on localhost to talk directly to the jvm. Also it gives consistent ipv6 and error handling. Two other incidental changes, that shouldn't matter: 1) python barrier tasks perform authentication immediately (rather than waiting for the BARRIER_FUNCTION indicator) 2) for `rdd._load_from_socket`, the timeout is only increased after authentication. Closes apache#22247 from squito/py_connection_refactor. Authored-by: Imran Rashid <[email protected]> Signed-off-by: hyukjinkwon <[email protected]> (cherry picked from commit 38391c9) (cherry picked from commit a2a54a5)
Willymontaz
pushed a commit
to criteo-forks/spark
that referenced
this pull request
Sep 27, 2019
This eliminates some duplication in the code to connect to a server on localhost to talk directly to the jvm. Also it gives consistent ipv6 and error handling. Two other incidental changes, that shouldn't matter: 1) python barrier tasks perform authentication immediately (rather than waiting for the BARRIER_FUNCTION indicator) 2) for `rdd._load_from_socket`, the timeout is only increased after authentication. Closes apache#22247 from squito/py_connection_refactor. Authored-by: Imran Rashid <[email protected]> Signed-off-by: hyukjinkwon <[email protected]> (cherry picked from commit 38391c9) (cherry picked from commit a2a54a5)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This eliminates some duplication in the code to connect to a server on localhost to talk directly to the jvm. Also it gives consistent ipv6 and error handling. Two other incidental changes, that shouldn't matter:
rdd._load_from_socket, the timeout is only increased after authentication.