On Wed, Nov 09, 2016 at 02:55:14PM +0200, Graham Inggs wrote: > Just to be clear: > - the build is always successful on some autobuilders > - the build randomly fails on the other autobuilders > > Is that right?
It is actually "better" than that: On the machine where it fails, it always fails (I used the term "random" because I had a mix of successful and failed logs). That's why I would like to investigate a little bit more before bothering you (as a last resort, I would offer ssh access to a machine where I could reproduce this, if I'm able to clone successfully the one on which it fails). Thanks.