@stephenw10 the net installer boots from USB, but installations don’t boot. If I use pure MBR it gets to the old F1/F2/F6 prompt but won’t go further.

@stephenw10 Both 24.03 and 24.11 are the same it after the rollback: 2.5 Gb physical interfaces on pfsense: only auto negotiates to 10/100Mb full duplex with TAP 1 Gb physical interfaces pfsense: auto negotiates to 1000Mb full duplex with TAP At this point, I am just a bit puzzled what changed I might have done. The TAP has been in use since like late 2021 and until like a month ago everything was working at gigabit on the same 2.5Gb physical ports on the pfsense.

@patient0 I do nightly incremental backups, and full backups every 2 weeks. I keep several full generations. The desktop data stripe is 8 x 1TB Samsung 860 SATA-3 SSDs. Boot drive is 2TB NVMe. The NAS uses 8 x WD 14TB HDDs, again, in RAID-Z2, so there is 84TB usable capacity.. The backups tasks use Acronis True Image with a WOL script to wake up the NAS, since I don't keep on 24/7 - 8 platters are loud and use a fair amount of power. The LAN is 10 gig so even a full desktop backup isn't too long.

Hmm, yeah so far I failed to replicate that but we are digging. I've no idea how it could ever show a cost like that....

It could be for some time after an upgrade while it rebuilds the cert database. Especially if you have a bunch of packages installed. It should finish doing that and drop back down to similar levels after that though. The dashboard consumes more CPU cycles in 24.11 because of the updated widget behaviour. That can be reverted if necessary.

Go to System > Boot Environments and remove old BEs.

Yes that error could be handled better. The 1100 hits it more than most because of the relatively low powered CPU. If you come from the dashboard to the update page the check may still be running in the background. But just refreshing the page should show it available as you found.

No, I do not have it activated. The image shows the CPU usage when the dashboard is not on screen. The high CPU usage is only when the dashboard is on screen.

That is the error I'd expect when trying to authorise against radius but I'd expect some other code when trying to start the service If it fails to start.

Just updated my NDI after adjusting NICs, and the above issue has been resolved automatically.

@terryzb I ignored the "Another instance of pfSense-upgrade is running." message and with 24.11 selected, ran option 13 again and this time it upgraded. Other than the CPU ranging between 70-100% and the GUI being very sluggish, it seems to be working.

Nice result! Yes the 3100 is EoL. It may or may not get 25.01 depending on how broken the build is. Additionally it can't run ZFS. The new upgrade process using ZFS BEs reduces downtime at upgrade a lot.

Possibly it is detected or initialised slightly differently in FreeBSD 15 and if that happens it might not create the ue0 interface. And that would cause pfSense to drop to the interfaces assign prompt due to an assigned missing NIC, eventually rebooting when the confirmation check timer expires. It would likely have worked if you had just unassigned the NIC. If you still need it you could reconnect it and see how it comes up in 24.11. Anyway glad you were able to get past it!

The reset button on the 1100 only resets the power. It doesn't reset the config. So it will still be running whatever it was last configured as.

@stephenw10 I turned on https and now my laptop works.

@TheGushi LOL that is a wonderful analogy.

@jimp So that cured the 4-5 high php-fpm items showing on the "top" while sitting on the dashboard and 100% CPU The only new observation after adding the system information widget back and adding the CPU usage item (which I had removed as a test) and then refreshing the entire dashboard was that everything was responsive except the CPU usage line item. It took a good almost 2 minutes to "retrieve" the value. which is now only a couple of points off (higher) where it would normally be which is <20% (really hard to tell when widgets are actually in the process of updating their data) Removing / adding and saving widgets is back to normal. back to 10-15 seconds instead of minutes+. Once it got past that initial retrieve of the CPU usage it seems to be updating in less time but hard to tell how long. I see no other dashboard issues, everything seems near "normal" php-fpm is no longer locked at the top of top and certainly not 4-5 copies of it. When it is running it is only on top briefly. There is really nothing stuck on top anymore it varies as normal. user is bouncing between 1 and 20% system between 3 and 10%. which seems more "normal" on this system Thanks for all the work you guys have put into this.