Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Chris Burr
    @chrisburr
    If anyone has ideas I'm happy to try them (or share a reproducer)
    Isuru Fernando
    @isuruf
    @chrisburr, if you didn't see that PR, adding -fplt or removing -fno-plt should fix this
    Chris Burr
    @chrisburr
    I didn't, that's great. Thanks for letting me know :tada:
    daryl herzmann
    @akrherz
    Seeing an issue over the past two days with Travis-CI libspatialite builds staling/hanging within mudane conda tasks for example, any known issues afoot or is this just standard Travis-CI struggles sometimes? TIA
    Matthias Diener
    @matthiasdiener
    I'm consistently seeing the following CI error on Travis/ppc64le:
    ####################################################################################
    Resource usage summary:
    Total time: 0:17:35.9
    CPU usage: sys=0:00:19.5, user=0:10:37.5
    Maximum memory usage observed: 716.6M
    Total disk usage observed (not including envs): 365.6K
    Validating outputs
    Cloning into '/tmp/tmpnwg8xke4_recipe/feedstock-outputs'...
    error: RPC failed; curl 56 OpenSSL SSL_read: Connection reset by peer, errno 104
    error: 2717 bytes of body are still expected
    fetch-pack: unexpected disconnect while reading sideband packet
    fatal: early EOF
    fatal: fetch-pack: invalid index-pack output
    Traceback (most recent call last):
      File "/opt/conda/bin/validate_recipe_outputs", line 11, in <module>
        sys.exit(main())
      File "/opt/conda/lib/python3.8/site-packages/click/core.py", line 1126, in __call__
        return self.main(*args, **kwargs)
      File "/opt/conda/lib/python3.8/site-packages/click/core.py", line 1051, in main
        rv = self.invoke(ctx)
      File "/opt/conda/lib/python3.8/site-packages/click/core.py", line 1393, in invoke
        return ctx.invoke(self.callback, **ctx.params)
      File "/opt/conda/lib/python3.8/site-packages/click/core.py", line 752, in invoke
        return __callback(*args, **kwargs)
      File "/opt/conda/lib/python3.8/site-packages/conda_forge_ci_setup/feedstock_outputs.py", line 176, in main
        results = is_valid_feedstock_output(feedstock_name, built_distributions)
      File "/opt/conda/lib/python3.8/site-packages/conda_forge_ci_setup/feedstock_outputs.py", line 126, in is_valid_feedstock_output
        subprocess.run(
      File "/opt/conda/lib/python3.8/subprocess.py", line 516, in run
        raise CalledProcessError(retcode, process.args,
    subprocess.CalledProcessError: Command '['git', 'clone', '--depth=1', 'https://github.com/conda-forge/feedstock-outputs.git', '/tmp/tmpnwg8xke4_recipe/feedstock-outputs']' returned non-zero exit status 128.
    The command "if [[ ${PLATFORM} =~ .*linux.* ]]; then ./.scripts/run_docker_build.sh; fi" exited with 1.
    3 replies
    Is this a known issue?
    jakirkham
    @jakirkham
    ^ @rpsene, could you please look into the issue above? We are not able to repro outside of Travis. So it looks like a Travis issue
    Isuru Fernando
    @isuruf
    jakirkham
    @jakirkham
    Sounds like there are still other sources of pain here ( conda-forge/conda-forge.github.io#1521 )
    jakirkham
    @jakirkham
    ^ @rpsene ?
    Rafael Sene
    @rpsene
    hey!
    sorry, I was offline yesterday.
    let me check it now
    jakirkham
    @jakirkham
    No worries. Thanks for taking a look :slight_smile:
    jakirkham
    @jakirkham
    @rpsene are there any updates from Travis on this issue?
    Rafael Sene
    @rpsene
    not yet; let me poke again
    is this caused by Let’s Ecnrypt issue?
    jakirkham
    @jakirkham

    We are seeing Error: HTTP 000 CONNECTION FAILED in jobs trying to download anything

    conda-forge/conda-forge.github.io#1521

    jakirkham
    @jakirkham
    @rpsene here's a recent build with this issue
    Gerardo Cervantes
    @gcervantes8
    I'm trying to install the package pyarrow to run with Red Hat Linux (PPC64le) using Conda, but I'm having strange issues when importing the package. I was wondering if anyone had some insight into the issues I'm facing or know of something similar. Details: https://issues.apache.org/jira/browse/ARROW-14323
    5 replies
    jakirkham
    @jakirkham
    @rpsene , sorry to bug you, but are there any updates on the HTTP connection issue noted above ( https://gitter.im/conda-forge-ppc64le/Lobby?at=616870b6d3ae402e474ac2f9 )?
    Axel Huebl
    @ax3l
    Same problem for me on my repos :(
    @rpsene ? :)
    Rafael Sene
    @rpsene
    hey folks
    from Travis: "I ran some tests and found no issues"
    1 reply
    Isuru Fernando
    @isuruf
    Still happening
    Rafael Sene
    @rpsene
    @isuruf sent you and email, request an account ar https://openpower.ic.unicamp.br/
    Isuru Fernando
    @isuruf
    Here's one half an hour ago
    Rafael Sene
    @rpsene
    at*
    Rafael Sene
    @rpsene
    if that works we can get a VM at IBm CLoud
    Isuru Fernando
    @isuruf
    @rpsene, is this different from the OSU OSL VMs? I've got one OSL VM and it works fine there
    just slow
    Above build ran on OSU OSL ppc64le server
    Rafael Sene
    @rpsene
    OSU and Minicloud (Unicamp) are Universites we at IBM work with to provide free access to Power.
    OSU in US and Unicamp in Brazil.
    Isuru Fernando
    @isuruf
    Ah, okay
    It works on OSU. I'll try Minicloud too
    Rafael Sene
    @rpsene
    OSU and Minicloud are KVM Based. At IBM Cloud you will use PowerVM, but as you are using Docker and Containers, this does not matter.
    Marcel Bargull
    @mbargull
    I'm not sure if the network issues you guys reported here early have the same cause or not, but @pkgw found a workaround for it in https://github.com/conda-forge/rust-feedstock/pull/94/commits/0b3add7c42b68e5787152732adab18d5262f8348 (reusing the host's network namespace for the build container).
    Isuru Fernando
    @isuruf
    Yes, it does. PRs are welcome for conda-smithy
    3 replies
    jakirkham
    @jakirkham
    @rpsene to summarize network access inside Docker containers is broken. Particularly we notice this when downloading from a GitHub URL. We have worked around this for now by adding --net=host to docker run on Travis, but this option is not recommend by Docker. It would be good if Travis took another look at what is causing this issue. Maybe there is a GitHub token that needs to get shared with Docker or something?
    Isuru Fernando
    @isuruf
    Anybody know what syscall 259 is on ppc64le? Seems like QEMU doesn't implement it
    Cheng H. Lee
    @chenghlee
    Isuru Fernando
    @isuruf
    ah, thanks
    Cheng H. Lee
    @chenghlee
    sets NUMA memory policy, so I wouldn't be surprised if it needs so memory features that qemu doesn't implement.
    Isuru Fernando
    @isuruf
    openblas is using mbind
    Cheng H. Lee
    @chenghlee
    Are you running full system emulation (qemu-system-ppc64le) or just the user static component?
    Isuru Fernando
    @isuruf
    user static component