More specifically: https://forum.pfsense.org/index.php?topic=113435.msg630835#msg630835

Thanks, I had found another thread that showed the same issue so I removed the PHP core file this morning following the instructions. I did download it first in case you wanted it. I thought I had updated the thread, but was multi tasking and probably forgot to post.. Usage is back down to 3% for now.

A good "adblock" shouldn't interfere with local sites ….. Consider dumping it - or add you local network to the "white list" (if it has one). I believe this is the best ad blocker on the market. It's always been an input selector in 2.3. Your adblocker changed/broke. Ok. The adblock is working properly. What was different was only updating the pfsense and google chrome to version 51.0.2704.79 (64-bit). I use linux.

So after some code diving and whatnot, I have arrived at the source of the problem, which in turn leads me to a new problem. On my system, I have System > Advanced > Miscellaneous > "Use memory file system for /tmp and /var" checked. /etc/resolv.conf is a symlink to /var/etc/resolv.conf, and because my /var is mounted from /dev/md1, when pkg goes to chroot into the other partition in order to apply the upgrade, there's nothing in /var/etc and the resolver can't figure out what DNS servers to use. So now I arrive at my other problem. I can't actually uncheck that value. I unset it, the machine reboots, and it's checked again when I log back in. According to my configuration file, it's not set at all. I have to assume that this is because I'm using the embedded build, since I'm running off a USB stick. I have to assume then that this problem affects all embedded users. I'm not sure how to proceed here, short of maybe putting in a hack into the upgrade script to copy the contents of /var into /tmp/nanobsd_upgrade/var before it starts the upgrade.

It depends on the driver support for the controller and the drive config. A common way to screw up an old IDE drive was to have it in the master position but set as slave, or set as slave with no master, either of which could cause it to not be seen right by an OS. Also check your BIOS for options that control how the controller behaves, it might be worth a tweak. Though honestly if that's really an old IDE disk it's probably not worth trusting. Disks are cheap, especially small ones, not worth the hassle of using a really old disk these days.

Ok I found my problem … better reboot when you change the proxy config  ;D

I double check. with my pfSense 07. who is running right now on 2.3.0 The lan interface is disable and all my vlan are working great.

Got space on the mobo to drop in a PCI or PCIe NIC?

What Phil said.

Dear Master  NOYB, Thanks for yours informations…

I got the solution from here https://forums.freebsd.org/threads/56524/ I had to disable SMP. *I use a very cheat and limited motherboard with literally no configurations in the bios.

The considerations are no different than with other hardware in general, except you don't have to worry about any kind of potential hardware compatibility issues, which takes away the biggest risk there. As always, pay attention to the release notes posted on the blog. Generally fine to upgrade to latest at this point with the exception of those on 2.2.x or earlier reliant on a package that's been deprecated in 2.3.x+. The upgrade process is the same.

Turning on legacy mode en the setup goes one step further I choose single user, kernel loads And oh, boot is going one step further : Enter full path name of shell or RETURN for /bin/sh: So I press enter : system hangs.  I type /bin/sh: <enter>Then an error : init: can't exec /bin/sh: for single user : no such file or directory. So I rebooted for the (forgot how many times but this time I did choose Multiple users and what a surprise, setup continues to the "graphical" setup. So case solved. Still weird that single user setup fails.</enter>

Ok so that has killed that process. The load is still high so it must be something else. last pid: 40086;  load averages:  0.50,  0.56,  0.51 What else can i try to track down the problem service. Thanks. Edit: This is what i am seeing now. ps auwwx | grep pkg root  48874  9.0  0.5  45172 10264  -  S    1:11PM    0:00.96 /usr/sbin/pkg search –raw-format json-compact pfSense-pkg- root  48739  0.0  0.4  45172  7796  -  S    1:11PM    0:00.00 /usr/sbin/pkg search --raw-format json-compact pfSense-pkg- Seems to come and go, so still not sure if thats the real service/process taking up the cpu time.

@churchi: Hi all, Ok so i have worked out what the issue was. Since my pfsense is in VMWare running as a VM on esxi, when i installed the last patch the system must have been rebooted which is normal. I had then rebooted the system a number of times to try and fix the issue but never powered down the VM. So tonight i powered down the VM and fired it back up and the interface stats are back on the dashboard. Weird one, but thought i would post so everyone knew the fix. Case solved. Cheers. A saint among men. Thanks for posting the solution. No DenverCoder9 here.

You do understand the iso is gzipped right, did you uncompress it to get your actual iso, and did you validate your checksum was valid?

When restoring a configuration and the interfaces don't match up. It prompts for a interface mismatch. This screen says select "Apply Changes" after fixing the mismatch. There is no apply changes button. The "Apply Changes" button got lost in among changes to the way info/warning boxes are displayed in 2.3 I raised Redmine issue https://redmine.pfsense.org/issues/6460 And Pull Request https://github.com/pfsense/pfsense/pull/2996

You can do it with nano, otherwise just backup the 2.2.6 config and upgrade. If it goes south for some reason, re-install 2.2.6 and restore the config.