@jimp said in Cannot uninstall nor upgrade squidGuard:

Hey, that's me! ... But I take no offense because it is a giant kludge. It does have some rare but specific useful purposes, but it gets used as a sledgehammer more often than not.

You... you... klutz 😁
Jokes aside sometimes I wish the package would be hidden in some "debug/expert" options only. I see that popping up far too often for my liking. Often the cultprits aren't even sure why they installed it or it's because of misunderstood problems (restarts of unbound because DNS errors etc.)

Can't we hide it from the UI and only allow it to be installed hands-on via shell/ssh? 😏
Or just throw a big red box of test above it so it's understood, that it's only a workaround but no "solution"? 😃

But as for the OP problem: is the "new" pkg tool somehow more sensible for running services? We had our share of problems after updating to 2.4.5p1 and afterwards, too and most of them are due to pkg deadlocking or stalling with packages, especially the big ones like FRR, Squid, pfBlocker?

As a side note: It would be nice to have a tool like service restart, that could use trigger/event conditions like other internal processes do e.g. WAN down/renewed IP, gateway fail, service xy restart, package (re/un)install, sync, whatever ;) That would be more in the way of the "shellcmd" package but it would make quite a few things easier to handle :) Especially a hook like "after any change written to config.xml" would be incredibly nice for doing some backup'py stuff.

These kinds of installation issues are quite rare.

I'm not sure we have enough information to precisely tell you the problem or how to fix it, but maybe we can start to work it out.

FreeBSD (and thus pfSense) can boot using UEFI since FreeBSD 10.1, however, depending on the BIOS/UEFI in your laptop it might help to enable CSM (Compatibility Support Module) to allow booting in legacy BIOS mode. CSM enabled should generally be the most safe option (my experience).

Regarding AHCI or IDE this is mostly to do with your SATA ports, but may affect USB drives aswell. AHCI mode should be default, IDE being legacy mode for very many years now.

Installation to USB drives is generally a bad idea, since the controllers and NAND chips on USB drives are not really geared for that kind of workload. You should REALLY consider installation to a small SSD or even a Harddrive. At least to work out if the issue is related to the one or more USB drives, that you have attempted to use.

So after hours of fun it truns out this is a bug. See https://forum.netgate.com/search?term=stuck%20after%20deinstall&in=titlesposts for more details.

In short you need to stop the squid services before the re-install can complete successfully!

Glad you have got it working. 👍

-Rico

Hi,

Contact support.
They give you the steps to update your SG-1100 so everything is fine again.

1: Boot offset could be your Bios (hw) clock being wrong.
Check & Correct

2: Investigate the NTP peers
Diagnostics -> Command prompt
Run the command : ntpq -p

3bbf1e64-255a-4308-b256-6276776cadfb-image.png

Paste output here

/Bingo

@reza-mnp said in unbound DNS Resolver Will Not Start:

Shared object "libevent-2.1.so.7" not found, required by "unbound-checkconf"

"libevent-2.1.so.7" is not /var/unbound/unbound.conf.

I saw this one https://forum.netgate.com/topic/154509/libevent-2-1-so-7-not-found and I can create that situation rather easily : by deleting that lib file.
Still, there is not enough info.

Hah, don't I feel stupid.

Thank you.

Wow.. Yeah.. Thanks! I think I'm in the market for some new hardware.. It was good to see all that in writing

Given that pfSense is a security product, keeping it up-to-date is critical to ensure that known security issues are addressed regularly. You don't have to jump right on every new release the moment it's out, but you should be getting things updated regularly.

There is more general advice in the upgrade guide to help ensure things go smoothly.

The default speed on the installer image and after is 115200, but you might have 9600 stored in your config.xml. Check under System > Advanced, Admin Access tab, at the bottom in the console section.

As for the speed in the APU BIOS, that should always have been 115200 as well, especially if you bought it from Netgate. To change that you'd have to go into the APU BIOS, directions for entering that should be printed at the start of the boot process.

@Cool_Corona hi! Another Dane here ;-)
Usually when I test up against 'Fiberby ApS' in speedtest.net I get several hundred Mbps.

I can see (from the page of the result ID of your screenshot) that we use the same provider (Nianet and I guess now it must be GlobalConnect for both of us). But of course we both might have some intermediary ISP in between (I have at least).

Please could I ask you to test up against 'Fiberby ApS' (and maybe rerun your original test again)? But only of course if you are currently running pfSense 2.4.5-p1 as I understand if you might needed to roll back to get full speed or do not have time for more tests atm.

I haven't updated yet to 2.4.5-p1 due to some preparations I have to make.

@revengineer Ahh, never did see that thread. Well I'm glad it's a known issue and I hope it gets fixed. Thanks for pointing that out and thanks for your help

db:0:kdb.enter.default> bt Tracing pid 12 tid 100266 td 0xfffff8000d85e620 kdb_enter() at kdb_enter+0x3b/frame 0xfffffe0c580f6740 vpanic() at vpanic+0x19b/frame 0xfffffe0c580f67a0 panic() at panic+0x43/frame 0xfffffe0c580f6800 trap_pfault() at trap_pfault/frame 0xfffffe0c580f6850 trap_pfault() at trap_pfault+0x49/frame 0xfffffe0c580f68b0 trap() at trap+0x29d/frame 0xfffffe0c580f69c0 calltrap() at calltrap+0x8/frame 0xfffffe0c580f69c0 --- trap 0xc, rip = 0xffffffff8051fdd0, rsp = 0xfffffe0c580f6a90, rbp = 0xfffffe0c580f6b20 --- bce_intr() at bce_intr+0x500/frame 0xfffffe0c580f6b20 intr_event_execute_handlers() at intr_event_execute_handlers+0xe9/frame 0xfffffe0c580f6b60 ithread_loop() at ithread_loop+0xe7/frame 0xfffffe0c580f6bb0 fork_exit() at fork_exit+0x83/frame 0xfffffe0c580f6bf0 fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe0c580f6bf0 --- trap 0, rip = 0, rsp = 0, rbp = 0 --- <6>bce1: promiscuous mode enabled bce2: Gigabit link up! Fatal trap 12: page fault while in kernel mode cpuid = 12; apic id = 20 fault virtual address = 0x10 fault code = supervisor read data, page not present instruction pointer = 0x20:0xffffffff8051fcf0 stack pointer = 0x28:0xfffffe0c582fba90 frame pointer = 0x28:0xfffffe0c582fbb20 code segment = base 0x0, limit 0xfffff, type 0x1b = DPL 0, pres 1, long 1, def32 0, gran 1 processor eflags = interrupt enabled, resume, IOPL = 0 current process = 12 (irq258: bce1) trap number = 12 panic: page fault cpuid = 12 KDB: enter: panic

Unfortunately there isn't much in the backtrace that would point to a specific cause, except it's crashing in the bce driver, apparently while processing an interrupt. Could be a driver issue, could be hardware.

@mauzilla said in can't upgrade from 2.4.3 to 2.4.5 with error message "Update origin security/php56-mcrypt -> ... failed":

Were you ever able to resolve this? I am hoping someone can still give an update on how to get this resolved? I am worried that if we restart the firewall

No luck so far, I haven't tried again since I saw people have issues on 2.4.5, I'll probably wait again until it's stable, btw my firewall reboot just fine, I'm running pfSense under Hyper-V so if anything goes wrong, I can restore the checkpoint.

reboot and run fsck

https://docs.netgate.com/pfsense/en/latest/hardware/troubleshooting-disk-check-errors-fsck.html

option 5 and F ffrom the menu

Once again the pfSense GUI Packet Manager got stuck after the most recent pfSense upgrade 2.4.5_1. Yet again it was pkg-static process which was stuck.

Executed:
killall pkg-static
pkg-static upgrade -f

Saved the day.