Upgrade from 10.4 stable to 11.0 releng fails

[Sven Vermeulen]

COMPAT_FREEBSD10 was indeed missing in my custom kernel config.

Now after a reboot the system continues, but then natd failed to start:

# /etc/rc.d/natd start
Starting natd.
natd: unknown interface name xl0
/etc/rc.d/natd: WARNING: failed to start natd

While xl0 is up and running:

# ifconfig xl0
xl0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
         options=82009<RXCSUM,VLAN_MTU,WOL_MAGIC,LINKSTATE>
         ether 00:50:da:e3:31:f6
         inet 82.94.48.144 netmask 0xffffffff broadcast 82.94.48.144 
vhid 255
         inet 10.0.0.150 netmask 0xffffff00 broadcast 10.0.0.255 vhid 3
         nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
         media: Ethernet autoselect (100baseTX <full-duplex>)
         status: active

Is this because of the new binaries (natd) are not yet installed?

Op 12-11-2018 om 19:56 schreef Eugene Grosbein:
> 13.11.2018 1:36, Sven Vermeulen wrote:
> 
>> I am busy trying to upgrade FreeBSD from 10.4 stable to 11.0 releng with make world.
>>
>> I did the following steps:
>>
>> 1) svn update (for /base/releng/11.0)
>> 2) make buildworld
>> 3) make buildkernel (using a custom kernel config)
>> 4) make installkernel (using a custom kernel config)
>>
>> installkernel gives some warnings with kldxref:
>>
>> kldxref /boot/kernel
>> kldxref: unknown metadata record 4 in file aic.ko
>> kldxref: unknown metadata record 4 in file atacard.ko
>> kldxref: unknown metadata record 4 in file atp.ko
>> ...
>>
>> I'm not sure if I can safely ignore these warnings.
> 
> Yes, you can. However, you better upgrade directly from 10.4 to 11.2 when using source upgrade path
> to save the time and troubles.
> 
>>
>> 5) reboot
>>
>> After the reboot the kernel seems to start fine but after that I get the following error:
>>
>> pid 22 (sh), uid 0: exited on signal 12
>> init: /bin/sh on /etc/rc terminated abnormally, going to single user mode
>> Enter full pathname of shell or RETURN for /bin/sh:
>>
>> When I run the command /bin/sh /etc/rc on the prompt I get the following error:
>>
>> pid 30 (sh), uid 0: exited on signal 12
>> Bad system call
>>
>> Any idea what is going wrong here?
> 
> signal 12 is SIGSYS meaning that your binaries use syscalls that kernel does not have,
> so you either still run old 10.4 kernel with new binaries or you run new custom 11.0 kernel
> built without options COMPAT_FREEBSD10 and old binaries.
> 
> There are many possible cases for this. For example, if you have degraded gmirror then
> you might installed new kernel and binaries to one part of the mirror only (due to its degraded state)
> but loader boots old kernel using another part of the mirror because it known nothing about gmirror
> and uses BIOS service to read plain disk (make it boot from another part forcebly then).
> 
> Or some another glitch, verify installed data carefully.
> 
> If you run new custom 11.0 kernel built without options COMPAT_FREEBSD10 and old binaries,
> then use loader to unload new kernel and load kernel.old before booting.
> 
> _______________________________________________
> freebsd-stable at freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-stable
> To unsubscribe, send any mail to "freebsd-stable-unsubscribe at freebsd.org"
>