It took me ages, but I found the difference between trunk and precommit
Precommits are executed on hadoop* machines. There is a single build at a
Trunks are executed on ubuntu* machines. There are two simultaneous builds.
For example, we have currently on ubuntu3:
ubuntu3 <https://builds.apache.org/computer/ubuntu3/> 1
I suppose we can have two hbase builds on the same machine.
Imho, that's why we have so many failures on trunk builds.
- Use hadoop* for trunk builds. That's the best option imho, they seem
cleaner then ubuntu* and less use.
- When building trunk, use a very low parallel test factor. May be 2.
Does anyone know how to do this?
Stack 2013-04-05, 18:30
Nicolas Liochon 2013-04-05, 18:36
Stack 2013-04-05, 18:52
Andrew Purtell 2013-04-05, 19:22
Andrew Purtell 2013-04-06, 00:03
Nicolas Liochon 2013-04-06, 11:38
Ted 2013-04-06, 12:33
Nicolas Liochon 2013-04-07, 12:52
Nicolas Liochon 2013-04-05, 16:11
Ted Yu 2013-04-05, 16:30
Nicolas Liochon 2013-04-05, 16:43
Stack 2013-04-05, 18:14
Ted Yu 2013-04-05, 16:15