Upgrade to 2.1.2: Stuck on 2.1

doktornotor

0/ I deleted nothing, I'm not even a moderator. Regardless, it has been deleted because you posted two idential posts in two different threads. Just do NOT do that.
1/ APU != ALIX. You seriously are confusing people here. Totally different HW!
2/ Last time I've heard, cross-arch upgrades are not exactly something supported. Why don't you just do a fresh install and restore the config?
3/ The APU is something far from tested as well.

mrsunfire

OK, the APU is from the same manufactor of the ALIX but called different. I will try a fresh install and come back later.

kaput

I'm seeing the same issue on an ALIX board. The upgrade goes through with no errors reported on screen/console, but when it reboots, it boots back into 2.1. The upgrade log has only the following that appears notable.

fdisk: invalid fdisk partition table found
bsdlabel: /dev/ad0s3: no valid label found
bsdlabel: /dev/ad0s3: no valid label found
bsdlabel: /dev/ad0s3: no valid label found
shutdown: [pid 75726]

sashka

@kaput are you on 4g image?

kaput

Yes.

sashka

Can someone test 4g image to see if it's not broken? I had no problems upgrading 2g images, but 4g gives problem.

jimp

Provide the whole upgrade log. The part being quoted isn't the problem.

We tested the 4GB version more than others, as it's the most common one in use.

kaput

http://pastebin.com/YX8vK7C1

jimp

Here's your problem:

gpart set -a active -i 1 ad0
gpart: table 'ad0' is corrupt: Operation not permitted

Try this:

gpart bootcode -b /boot/mbr ad0

Then try the upgrade again.

kaput

Thanks, jimp. Missed that perm issue in the log.

Unfortunately, running "gpart bootcode -b /boot/mbr ad0" also results in the exact same error, noting that ad0 is corrupt.

prohidium

I have the exact same issue with same error in my upgrade log on a 2g install on an Alix box.

I also had the issue on a 4g install but I opted to flash a new image and swap cards. The 2g box though is 4 hours round trip away so I would love to fix the issue and update remotely.

I also had another 4g box that updated just fine. So, 2 out of 3 are broken and would not update with autoupdate or manual update.

jimp

Other things to try:

Set the boot slice on Diagnostics > NanoBSD, see if it actually changes.

or

gpart recover ad0

or

sysctl kern.geom.debugflags=16
boot0cfg -v -s 1 ad0
fdisk -a /dev/ad0
sysctl kern.geom.debugflags=0

(For fdisk make sure to specify the values it asks and press 'y' to confirm!)

We have yet to be able to reproduce this effectively, but several people have hit it. It would be helpful to know some more about the original setup that failed. Was it your own CF or did it come with the device? Had it ever been upgraded before, and if so how many times? What was its original version of pfSense on that CF card before upgrading?

sashka

Slice does not change; nor I can copy ad0s2 -> ad0s1.

kaput

 gpart bootcode -b /boot/mbr ad0
gpart: table 'ad0' is corrupt: Operation not permitted

gpart recover ad0
gpart: recovering 'ad0' failed: Function not implemented

sashka


fdisk -a /dev/ad0

I get an Intput/Output error.

This is compact flash card. It was installed as 2.0 and then upgraded to 2.0.3 and later 2.1.

robi

In my case I had pfSense nanobsd v1.2.3 upgraded to v2.0, that one upgraded to v2.1 - all of them upgraded OK. Now when I wanted to upgrade to v2.1.2, It seems corrupted. I'm preparing a fresly flashed CF card with config already put in.

me too

I am having the same issue. Hopefully my situation will help shed some light.

I have upgraded at least two other NanoBSD 4G (ALIX) system without issue.
Today I am upgrading three identical boxes - ALIX boards running 4G CF card
One existing install was already 2.1 and I successfully upgraded to 2.1.2
The other two were at 2.0.3 and I successfully upgraded to 2.1, but they are both now stuck on 2.1 after auto-upgrade and manual upgrade attempts.

robi

Also a side note, I'm not on Alix, but on a Jetway JNF99 with 5 intel nics, Atom CPU and 4GB of RAM, i386 build. The new CF card will carry a 64-bit version this time.

jimp

If you're getting an "i/o error" or "device not configured", reboot and try again. At least for the "i/o error" it seems to lean more toward maybe a bad CF but it's still tough to say for sure.

Also worth a try:

sysctl kern.geom.debugflags=16
fdisk -B -b /boot/boot0 ad0

mjohnson

I'm in the same boat with 40 plus devices scattered across 4 provinces. I actually haven't managed to get a single one to upgrade except the 3 in our local offices that I performed a clean install on, 2G and 4G, current versions 2.03 and 2.1 on all remotes attempting auto upgrade and manual firmware uploads. Not sure what to do, since I don't want to risk chopping them up and losing remote connectivity. I don't own a helicopter to get around that quickly ;D

All Alix boards. Errors in the upgrade log

fdisk: invalid fdisk partition table found
bsdlabel: /dev/ad0s3: no valid label found
bsdlabel: /dev/ad0s3: no valid label found
bsdlabel: /dev/ad0s3: no valid label found
tar: Failed to set default locale
tar: Failed to set default locale
shutdown: [pid 24752]