On 02/18/2013 10:09 AM, Guido Günther wrote:
On Sat, Feb 16, 2013 at 05:53:05PM -0500, Laine Stump wrote:
On 02/16/2013 12:20 AM, Doug Goldstein wrote:
The following error bisect's down to this commit when running out of my local checkout for testing.
2013-02-16 05:16:55.102+0000: 29992: error : virCommandWait:2270 : internal error Child process (LC_ALL=C LD_LIBRARY_PATH=/home/cardoe/work/libvirt/src/.libs PATH=/usr/local/bin:/usr/bin:/bin:/opt/bin:/usr/x86_64-pc-linux-gnu/gcc-bin/4.6.3:/usr/games/bin HOME=/home/cardoe USER=cardoe LOGNAME=cardoe /usr/bin/qemu-kvm -help) unexpected exit status 1: libvir: error : internal error cannot apply process capabilities -1
Ugh. Can you manage to get that trapped in gdb and find out the value of uid, gid, and capBits, as well as whether it is failing on the first call to capng_apply() or the second (they both have the same error messsage. (Whatever happened to the function name/line number that used to be logged with the error messages?) I wonder if perhaps on debian it's failing the capng_apply() call that happens after the uid is changed... It's uid = 0, gid = 0 (as can be seen when running with LIBVIRT_DEBUG=1) . See 20130217173308.GA11314@bogon.sigxcpu.org for a proposed fix.
Ah, good. I see that one now (after searching through my google apps spam folder.. grumble grumble.) I was suspicious there might be some fallout due to no longer treating uid=0 as "ignore/don't change", but didn't think to try session mode libvirtd. Thanks for figuring it out.