On Fri, Apr 23, 2010 at 01:14:22PM +0200, [email protected] wrote: > Hi list, > > I found some traces of this 'issue' (if it actually is one, no idea), e.g. > here: > > http://www.pubbs.net/openbsd/200911/51706/ > > host A: > > # cat /etc/hostname.carp1 > > inet 10.0.0.5 255.0.0.0 NONE vhid 2 > > host B: > > # cat /etc/hostname.carp1 > > inet 10.0.0.5 255.0.0.0 NONE vhid 2 advskew 100 > > On both hosts: > > # sysctl |grep carp > > net.inet.carp.allow=1 > net.inet.carp.preempt=1 > net.inet.carp.log=2 > > # dmesg |head > OpenBSD 4.6 (GENERIC.MP) #89: Thu Jul 9 21:32:39 MDT 2009 > [email protected]:/usr/src/sys/arch/i386/compile/GENERIC.MP > cpu0: Geode(TM) Integrated Processor by AMD PCS ("AuthenticAMD" 586-class) > 499 MHz > cpu0: FPU,DE,PSE,TSC,MSR,CX8,SEP,PGE,CMOV,CFLUSH,MMX > real mem = 268009472 (255MB) > avail mem = 250314752 (238MB) > mainbus0 at root > bios0 at mainbus0: AT/286+ BIOS, date 12/10/07, BIOS32 rev. 0 @ 0xfceb2 > pcibios0 at bios0: rev 2.1 @ 0xf0000/0x10000 > > (ALIX boards by PC Engines.) > > However, if host A is MASTER (the 'usual' setup) and then reboots, host B > becomes MASTER -- perfect. However, if host A is back online again, it does > NOT become MASTER again. Host B stays MASTER (either until its rebooted or > the carp or underlying interfaces are shut down). > > Is this a known behaviour? Is it wanted? Thanks for shedding light on this.
Yes, this is known. "your" mail archive does not list the followup, but here it is: http://marc.info/?l=openbsd-misc&m=125959839221255&w=2 In 4.6, if you do not use the workaround, it can take several minutes for the rebooted master to become master again (depending on your max states setting in pf.conf), but it should do so in the end. This has been fixed in 4.7. -Otto
