Skip to content

Lock grpcio version#811

Merged
mc-nv merged 1 commit intor24.11from
mchornyi-24.11-grpcio
Nov 21, 2024
Merged

Lock grpcio version#811
mc-nv merged 1 commit intor24.11from
mchornyi-24.11-grpcio

Conversation

@mc-nv
Copy link
Copy Markdown
Contributor

@mc-nv mc-nv commented Nov 20, 2024

Intro:

Without this upper version limit model we getting issue where Model Analyzer is unable to reach grpc endpoint.

@kthui
Copy link
Copy Markdown
Contributor

kthui commented Nov 20, 2024

Hi @krishung5 @rmccorm4, do you have any concern if the gRPC version on the client is newer than the one on the server?

The server is currently on 1.54.3, and this change will set the version on the client to 1.67.1

@rmccorm4
Copy link
Copy Markdown
Contributor

rmccorm4 commented Nov 20, 2024

Hi @krishung5 @rmccorm4, do you have any concern if the gRPC version on the client is newer than the one on the server?

The server is currently on 1.54.3, and this change will set the version on the client to 1.67.1

@kthui The current client version specification before this change (>=1.63.0) is already higher than the server version you pointed to, and hasn't given us any issues I'm aware of, so I have no issue with this change.

However, @mc-nv can you add details on the reason/context in the Description of these types of PRs so in the future looking backwards we can better understand why this change was made?

Also linking related PR: triton-inference-server/model_analyzer#946

@kthui kthui self-requested a review November 20, 2024 22:37
@mc-nv mc-nv marked this pull request as ready for review November 21, 2024 03:15
@mc-nv mc-nv merged commit a77904d into r24.11 Nov 21, 2024
@mc-nv mc-nv deleted the mchornyi-24.11-grpcio branch November 21, 2024 20:34
mc-nv added a commit that referenced this pull request Nov 22, 2024
mc-nv added a commit that referenced this pull request Nov 26, 2024
* Build Updates for Ubuntu24.04 (#802) (#805)

* update cmake version

* include patch version in cmake install

Co-authored-by: Anant Sharma <anants@nvidia.com>

* Hotfix for java (#808)

* Prevent failure if cache is empty (#809)

* Lock grpcio version (#811)

* fix: Partial fix for L0_https (#810)

* Fix L0_https--base SSL issue (#813)

* Upgrade geventhttpclient version to 2.3.2

* Upgrade geventhttpclient to latest version

---------

Co-authored-by: Anant Sharma <anants@nvidia.com>
Co-authored-by: Yingge He <157551214+yinggeh@users.noreply.github.com>
@springmeyer
Copy link
Copy Markdown

Found this issue trying to understand why grpcio is being restricted to <1.68. Are there plans to allow the latest grpcio release (1.76.0)?

Asking because I'm noticing that the latest grpcio release (1.76.0) supports binary wheels for python 3.14 but the 1.68 release does not. I can likely workaround this, but currently hitting this compile error (hence why I'm curious about supporting more recent grpcio to make this a non-issue):

 In file included from /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk/usr/include/stdio.h:61:
    | /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk/usr/include/_stdio.h:318:7: error: expected identifier or '('
    |   318 | FILE    *fdopen(int, const char *) __DARWIN_ALIAS_STARTING(__MAC_10_6, __IPHONE_2_0, __DARWIN_ALIAS(fdopen));
    |       |          ^
    | third_party/zlib/zutil.h:147:33: note: expanded from macro 'fdopen'
    |   147 | #        define fdopen(fd,mode) NULL /* No fdopen() */
    |       |                                 ^
    | /Library/Developer/CommandLineTools/usr/lib/clang/17/include/__stddef_null.h:26:16: note: expanded from macro 'NULL'
    |    26 | #define NULL ((void*)0)
    |       |                ^
    | In file included from third_party/zlib/zutil.c:10:
    | In file included from third_party/zlib/gzguts.h:20:
    | In file included from /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk/usr/include/stdio.h:61:
    | /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk/usr/include/_stdio.h:318:7: error: expected ')'
    | third_party/zlib/zutil.h:147:33: note: expanded from macro 'fdopen'
    |   147 | #        define fdopen(fd,mode) NULL /* No fdopen() */
    |       |                                 ^

@springmeyer
Copy link
Copy Markdown

Without this upper version limit model we getting issue where Model Analyzer is unable to reach grpc endpoint.

I'm hoping that underlying issue has now been resolved? Here is a PR to loosen requirements at #862, that should work if my assumption holds :)

@ozancaglayan
Copy link
Copy Markdown

Hello. This pinning avoids the use of Python 3.14. Can you please fix this? Thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Development

Successfully merging this pull request may close these issues.

7 participants