conda-smithy: AppVeyor VS2008 slow startup/memory issues
Seems there is a rather strange issue happening on the VS2008 builds unrelated to the recipe here. They seem to be taking a really long time to run and may be running out of memory as this build shows. Not sure if this is a conda problem (infinite recursion and so on) or a VM memory problem.
xref: https://ci.appveyor.com/project/conda-forge/iteration-utilities-feedstock/build/1.0.10/job/m42dxwpke0i6e5pt#L128 xref: https://ci.appveyor.com/project/conda-forge/conda-smithy-feedstock/build/1.0.89/job/1xvlseif9w9se8tr#L120
About this issue
- Original URL
- State: closed
- Created 8 years ago
- Comments: 21 (16 by maintainers)
This appears to have resurfaced.
xref: https://ci.appveyor.com/project/conda-forge/libssh2-feedstock/build/1.0.4/job/xcra9t5bevmcj3d8#L120
Maybe. Really all we can do is systematically eliminate problems until we are left with the source of the issue. Right now it looks like
ruamel_yamlis part of the problem. So eliminating that variable should help.Also @MSeifert04 I stopped and restarted that build. Not because it is occupying the queue, but because it looks like it was going to timeout anyways (AppVeyor has a 1hr limit). As all the other builds passed and deployed, this should make sure that one build should too. Hope that is ok.
@jakirkham Here’s the copy you asked for (https://github.com/conda-forge/iteration_utilities-feedstock/pull/3):
See for example (today):
2 days ago:
4 days ago:
9 days ago:
And those are just those I’ve looked into. So far a 100% hit rate.
11 days ago (>>>no delay<<<):
So probably at some point 10 or 11 days ago something changed/broke.
FWIW, I am not seeing delays in python 2.7 jobs in other (non-conda-forge-related) appveyor jobs…