Running 2.2 for about 24 hours here.  I played it safe and installed to a spare SSD on a second machine, swapped drives and restored the 2.1.5 configuration backup.  Only running a few packages, snort, nut, avahi and they reinstalled fine after the restore reboot.

The only minor thing I've noticed is that the gateway dashboard widget is showing impossibly low RTT values, on average about 5ms when it should really be about 25ms, which is what 2.1.5 was showing (ping shows about the same, mid 20's).

Thanks to those involved in this release!

Looks like the same issue as this:
https://forum.pfsense.org/index.php?topic=87290.0
If you post in that thread perhaps the two of you can find something you have configured similarly.

Steve

I'll have to take a look when I can shut it down again..  It had been working without issue on 2.1.5 with the current bios settings.. nothing was touched on the motherboard between 2.1.5 and 2.2.

The motherboard is pretty ancient, maybe 5 years old at this point.

Success.  While I was not able to run pfsense 2.1.x on Soekris net5501 and Alix2d13 hardware platforms after restoring my configurations without kernel panics … Now that pfsense 2.2 Release is available ... I have successfully updated both hardware platforms from pfsense 2.0.3 Release to 2.2 Release and all is well after restoring my 2.0.3 configuration.  I'm happy.  I've learned to be patient in life as in the end it's very worthwhile.

p.s. for OVPN I had to disable a deprecated option for tls-remote and optionally add an option for auth-nocache

-- Anthony Tzouris

@TieT:

Is it also fully enabled on the WAN or LAN ?
Services - snort - interfaces - the little arrow is green ?

WAN Enabled.

Hopefully the package reinstall has finished by now. If not, then go to Diagnostics->Backup/Restore and press the "Clear Package Lock" button. Then look in the logs for error messages about what went wrong during the package install. Then (re)install your packages.

The bug report is the way to go… last i knew there was no real maintainer for vnstat2.. Its been worked on by different people.. I was going going to fix the paths this morning but someone beat me to it :-)  Other packages have the same issue with the paths being wrong

Yup i'm glad pfsense finally moved to freebsd 10.1 ! that freebsd 8 was so legacy it made the project look "outdated" ;-)

I was able to upgrade one of my boxes just fine and the installer re-installed the packages automagically and all works except unbound … but, i hope somebody replies to my "unbound" post soon :)

I'm just excited that i get to use VMXNET3 now since Freebsd 10 supports it ( on my home machine that sometimes has to deal with 1GBPS routing between VLANs ), other than that ... i guess a "multi core" pf with freebsd 10 is good too !

Update:  So I rebooted again and now ssh does auto start?

Not sure if it fixed itself, or if running the commands below helped.  Found them in a post about ssh problems with the prerelease 2.2.  There was a quote from the script that I assume is part of the clean install process.

https://forum.pfsense.org/index.php?topic=83333.msg456428#msg456428

/usr/bin/ssh-keygen -t ecdsa -N '' -f /etc/ssh/ssh_host_ecdsa_key
/usr/bin/ssh-keygen -t ed25519 -N '' -f /etc/ssh/ssh_host_ed25519_key

So at this point I think am working.

No worries, we've all done it.  ;)

Steve

squidguard is broken in 2.2

this may help for squidguard but squid2.7 i believe is broken from other reports on the forum. try squid3?

https://redmine.pfsense.org/issues/4216

@kdillen:

But I do know that some things get corrupt when importing a 2.1.x file into 2.2

In my case I had the following problems with corrupt files:

ssh host key files unbound: root.key  (auto-trust-anchor-file)

That's absolutely not true. Those two things don't even exist in the config.

It's always safe to restore the full config from an older version to a newer version (but going backwards, say restoring 2.2 to 2.1, is not possible).

Trying to snip out pieces of an old config backup and paste them into a new config isn't a good idea unless you really know what you're doing, as those pieces may require config upgrades when migrating to a new version that they won't get if you're pulling in partial config bits.

2.2 omits localhost from resolv.conf if you aren't binding on it, to prevent that from being an issue. That'll cause a timeout in earlier versions if you have the system configured to use localhost as a DNS server, but aren't actually binding it there. Make sure if you're selectively binding DNS Forwarder, localhost is one of the selections unless you disable its use under System>General Setup.

Ha!

I'm all about Easy Editing. I've been stuck in some obscure vi mode too many times.  ::)

Steve

Thanks for the suggestion to restore from the config backup.  :)
I wasn't aware I could selectively restore stuff, but then chose to restore only the rrd data. I was logged in through ssh during the whole process and it was actually fun to watch it upload the data, incorporate it into config.xml, then write out the /var/db/rrd, and remove it from the config file. Before rebooting, it saved the /var/db/rrd data into rrd.tgz, and unpacked it again after the reboot. So, my data is back, with only a short missing piece during the upgrade process.

Roman

I see how that isn't the expected behaviour from your point of view.  :-\  A box showing the current update URL would perhaps help there.
That's the correct behaviour though, there's nothing wrong with your install.

Steve