Closed Lnaden closed 5 years ago
Thanks! I'm a bit worried about corrupting our working images, but if this rebuild is unsuccessful, it won't disrupt builds until I rebuild the CUDA images, so I think this is OK to try!
Not sure how it would corrupt? This update would not change how the downstream's are constructed are they? Unless the source of the downloaded material changed, this should build the same way.
You could make another spinoff builds just in case? Or make local copies of the current images in case you have to push a backup
Not sure how it would corrupt? This update would not change how the downstream's are constructed are they? Unless the source of the downloaded material changed, this should build the same way.
That's what I was worried about---some upstream packages might have changed!
It broke already: https://hub.docker.com/r/jchodera/omnia-linux-anvil/builds/bzebtyryendlkbtnvde5ul9/
Build failed: ADD failed: Get https://tycho.usno.navy.mil/timer.html: x509: certificate signed by unknown authority
Any ideas?
See #14
This should help advance omnia-md/conda-recipes#911 by pushing a change to the base docker image for the conda-build version workaround.
Once this is merged in @jchodera , all you should have to do is have this build on your Docker account (automatic?), then re-trigger all the various
cuda
images.