On Wednesday 25 November 2015 11:55:24 Miroslav Lichvar wrote: > On Wed, Nov 25, 2015 at 10:45:12AM +0000, Peter Humphrey wrote: > > On Wednesday 25 November 2015 10:06:22 Miroslav Lichvar wrote: > > > Hm, bugs that disappear when trying to debug them are the worst kind. > > > > > > I'd suggest to ask the OpenRC maintainers if they have any idea what > > > could be wrong here or have some hints on how to debug it. > > > > Maybe I'm going round in circles here, but I've just established that > > setting rtcsync in chrony.conf causes the error, while setting rtcfile > > does not. That leads me to suspect the kernel config again, so I'll go > > and make some coffee... > > I suspect that just changes the timing between fork and exit of the > foreground chronyd process, which somehow avoids a race condition. You > could try to put "usleep(100000);" after the go_daemon() call in > main.c in the chrony source code and see if that doesn't "fix" the > problem too.
You're right - it does. (On this Gentoo box I had to unpack the source, make a copy, edit the copy, diff the two to make a patch, copy /usr/portage/net-misc/chrony/ to a local overlay, copy the 2.2.ebuild to 2.2-r1.ebuild in the overlay, add to it an epatch call to apply the patch, create a new manifest - then emerge chrony. Simple! The box is now running an unofficial version 2.2-r1 of chrony.) Of course, the simplest thing for me would be just to change rtcsync to rtcfile in chrony.conf, delete all that -r1 stuff and hope no-one else stumbles into the same problem. That might not suit you though. -- Rgds Peter -- To unsubscribe email chrony-users-requ...@chrony.tuxfamily.org with "unsubscribe" in the subject. For help email chrony-users-requ...@chrony.tuxfamily.org with "help" in the subject. Trouble? Email listmas...@chrony.tuxfamily.org.