**** BEGIN LOGGING AT Sun Mar 30 02:59:59 2014 Mar 30 12:04:12 any idea why fetcher issues aren't shown in cooker log anymore? Mar 30 17:39:51 oha, der harry is da Mar 30 20:48:55 RP: I figured out how to reproduce my sstate reuse problem finally. In build A, build from scratch. In build B, build from build A's sstate-cache using SSTATE_MIRRORS. In build C, either use build B's SSTATE_DIR directly, or via SSTATE_MIRRORS. The sstate-cache produced by the build which was pulling from an sstate mirror is incomplete, it doesn't match up with what build A produced, and that incomplete cache results in the rebuilds. The same Mar 30 20:48:56 effect is seen by altering build B to not use SSTATE_MIRRORS but then rebuild from its own sstate-cache. Mar 30 20:49:27 RP: Still digging into the exact delta between the working and broken sstate cache directories. Mar 30 21:59:44 kergoth: interesting. I'll be interested to understand that when you get to the bottom of it... Mar 30 22:18:12 I'm wondering if this is #4102 I'm seeing... I have a kernel-x.y.z ipk and a kernel-x.y.z-rt ipk under deploy/. Even though my local.conf points out the rt-kernel as the preffered version, the non-rt kernel package gets installed in the image, according to buildhistory Mar 30 22:24:32 s/local.conf/distro.conf Mar 30 22:29:53 and s/preferred version/preferred provider Mar 30 22:59:19 kergoth: I just found an sstate sig issue affecting my local builds. Its unlikely its the one you're chasing but probably worth a quick check **** ENDING LOGGING AT Mon Mar 31 03:00:00 2014