Power outage during OpenBSD upgrade from 6.7 to 6.8

[cortez_h]

Hello,
i'm looking for help with a specific problem I encountered due to power outage during OpenBSD upgrade from 6.7 to 6.8

At the very beginning I need to point out that I'm have somewhat limited experience with Linux/Unix but managed to install the system, set up the neccesary httpd/ftp and ssh for the purpose I needed, and kept the system upgraded and patched without problems so far. Now I'm not confident what to do at the moment.

The system has been installed for over a year on an ancient hardware running 6.7 version just fine.

I followed the upgrade using sysupgrade command.
The /usr had over 1,3 GB free space, I've checked that before.
The upgrade packages were downloaded, and the script started the unpacking and upgrade processs.

After reboot, at some point, presumably during unpacking some userland - the system shut down probably due to power outage.

After recoupling the power it booted into 6.8 as it says:

cortez@ttyp0:~$ uname -a
OpenBSD satellite.home 6.8 GENERIC#427 i386

however it fails relinking the kernel

cortez@ttyp0:~$ cat /usr/share/relink/kernel/GENERIC/relink.log
(SHA256) /bsd: FAILED

Failed to verify /bsd's checksum, therefore a randomly linked kernel (KARL)
is not being built. KARL can be re-enabled for next boot by issuing as root:

sha256 -h /var/db/kernel.SHA256 /bsd

/home/_sysupgrade is rather empty

cortez@ttyp0:~$ ls -l /home/_sysupgrade/
total 0
-rw-r--r-- 1 root wheel 0 Apr 27 09:10 SHA256.sig

It seems that no data that was previously downloaded is present on the hdd, as the amount of freespace on /usr is almost without change to what was before running the sysupgrade

cortez@ttyp0:~$ df -h /usr
Filesystem Size Used Avail Capacity Mounted on
/dev/wd0d 2.9G 1.0G 1.7G 38% /usr

However veryfying else seems to work fine which means at least that I can boot up, login and main services such as networking, httpd, ftp, sshd, and ntpd work as before.

Now I'm unsure in what exactly state the system is, which means I'm unsure if it upgraded correctly as there was forced break during the process.

So my question is - what should I do to make sure that the system upgraded or not?
Where can I check any logs of what the sysupdate has done and where it stopped?
How to sanitize the situation? I wouldn't like to mess up more so before I start to do anything, I'd like to kindly ask for your assistance and some help.
Thank you in advance.

[jggimi]

Hello, and welcome!

During boot up, the kernel gets relinked so that with the next boot, module locations shift randomly, which prevent several classes of attacks.

A unique kernel is linked such that the startup assembly code is kept in the same place, followed by randomly-sized gapping, followed by all the other .o files randomly re-organized. As a result the distances between functions and variables are entirely new. An info leak of a pointer will not disclose other pointers or objects. This may also help reduce gadgets on variable-sized architectures, because polymorphism in the instruction stream is damaged by nested offsets changing.

The OS will stop further relinking if the relinking chain is ever broken, such as a reboot or power failure during the relinking process. The repair is simple, as was noted in the error message you posted above:

 # sha256 -h /var/db/kernel.SHA256 /bsd

[cortez_h]

Should that be enough? Is there any way to check if there arn't any files missing or corrupted?
And what about sysmerge? Should I use this command?

[TronDD]

Check root's mailbox. The upgrade process should email you a bunch of messages if it completed.

[cortez_h]

Unfortunately the root mailbox has not been touched since installation of the system

root@ttyp0:~# ls -l ./mbox
-rw------- 1 root wheel 4765 Jul 22 2020 ./mbox

No mail in other superuser account.

[TronDD]

The mailbox is /var/mail/root

I would rerun the upgrade by booting the bsd.rd and doing it the old fashioned way. If the install didn't finish, who knows what userland files are old. Installboot wouldn't have been run, etc.

[cortez_h]

/var/mail/root is nonexistent

root@ttyp0:/var/mail# ls -l
total 68
-rw------- 1 cortez cortez 33442 Apr 29 02:37 cortez

So, If I understand correctly, your suggestion is to rerun the upgrade again
running the interactive upgrade as described here: https://www.openbsd.org/faq/upgrade68.html

Which if I get it correctly means that I should manually download bsd.rd, then boot it and follow the upgrade path.

Sounds reasonable

[TronDD]

Yes. Someone correct me if I am wrong, but I think sysupgrade will believe you are already upgraded and won't run again so you'll have to do it with the bsd.rd

I suppose you can try sysupgrade again but we're right on the edge of the 6.9 release and I suppose it's possible that it might try to upgrade you to 6.9 unexpectedly.

[cortez_h]

Thank you.
I've gone through the process as advised, and it went smoothly from the beginning to the end.
It has als sent me the e-mails about the sysmerge and fristrun so I suppose it's ok now.
Thank you.