Poudriere bulk builds sometimes take two attempts

K

kjpetrie

This morning three packages failed to build and another four were skipped. The failed packages were devel/icu, devel/pcre2 and www/py-requests-toolbelt. I ran poudriere bulk manually and the seven packages were all built on the second run. This is not the first time failed builds have been corrected by a second run.

Would I be right in thinking this suggests poudriere had got the order for building these packages wrong and that this could reflect missing information on build requirements in the ports concerned?
 
Post the failing build logs so we can see why they failed. Please use something like misc/pastebinit as these logs are typically rather large.
 
c++: error: unable to execute command: Killed
…
build of devel/icu | icu-65.1,1 ended at Thu Oct 24 10:01:13 BST 2019
Click to expand...

Child process pid=94630 terminated abnormally: Killed
…
build of www/py-requests-toolbelt | py36-requests-toolbelt-0.8.0_1 ended at Thu Oct 24 10:01:14 BST 2019
Click to expand...

Not enough RAM?
 
This VM has over 4GB RAM, but with 8 processors building in parallel maybe that's not enough for every eventuality. It's possible, I suppose. The simultaneous failure of two builds could suggest a resource problem. It certainly suggests an external cause rather than a problem in the ports themselves, come to think of it.
 
You must log in or register to reply here.
Back
Top