I'm a little confused about the recent OpenZFS release issues

OpenZFS 2.2.2 is out to fix data corruption bugs

but FreeBSD 14.0-RELEASE-p1 contains patches to OpenZFS 2.2.0?

makes sense, cherrypicked against what was tested

veg: No, FreeBSD doesn't have 2.2.2. The issues FreeBSD had were in 2.2.0. This was fixed in 2.2.1 but another whopper bug was introduced, so 2.2.2 fixes both for Linux, while not yet addressing the still-newer LUKS issue.

i have options GDB set (13.2) but debug.kdb.available does not contain any entries. my intention is to use bhyve -G, is there anything else i need to do to be able to set debug.kdb.current=gdb?

hello

I am having merge conflicts and trying to upgrade to 14.0 from 13.2

is there any advice before I just run the backup to reverse this entire nonsense?

ls

exit

sixpiece: I had a few issues, and decided to save my config/data offsite and just install a fresh version of 14-RELEASE and then made sure to upgrade to the p1 that is out now

sixpiece: no issues so far, running strong; 14-release-p1

yes I should stop trying to be on the forefront so much and like wait 6 months before I install it I am trying something new but I will just reverse it if it doesn't work I guess

13.2 was also running very well, I'd say unless you need a new stable version of the utility that's specifically available in 14.0, just relax a bit until it's updated some, yes

yes I will try it again in 6 months

thank you

np, have a good night and thanks for using freebsd :)

Hey there! I'm finding that my laptop with FreeBSD freezes up for a short bit (2-3 seconds) every now and then, the graphics freeze but the mouse pointer doesn't, so it kinda sounds to me like a system call is taking a long time to complete. I don't see any messages in dmesg or on tty0, any ideas on how I could debug this?

I use ZFS on GELI with this laptop, with Intel graphics. I've upgraded to 14.0-RELEASE a few days ago, but I had this same behavior on 13

i suspect in dmesg you will see something about GPU hang

angry_vincent: I didn't see anything. I waited for it to happen, then switched to a TTY with an old dmesg listing, asked for a new listing, and nothing new appeared there

voy4g3r2: "tried sudo su toor" ?

voy4g3r2: tried "sudo su toor" ?

voy4g3r2: reboot machine, boot to single-user mode, set root password, switch to multi-user and bring up services

voy4g3r2, Reboot, at the boot menu select E to select a boot Environment, select A to select the previous active Boot Environment, log in normally, bectl mount default /mnt, vipw -d /mnt/etc, fix the broken password file, bectl umount default, shutdown -r now, boot normally, log in normally. Profit!

If you want to use a different editor, like I do, then at the vipw step specify a different editor with "env EDITOR=emacs vipw -d /mnt/etc" or use whichever editor you prefer.

The use for "env" there is needed for when csh is the command line shell (such as the previous FreeBSD 13) but env works with any shell and so is often documented to use for portability.

When booting the previous Boot Environment you are logging into a file system clone of the snapshot freebsd-update created just before upgrading. It's the previous OS from before the problem.

Also this is probably obvious but while using "vipw -d /mnt/etc" to fix the passwd file the previous file with the previous entries are available to reference at the /etc/master.passwd file. If things are really mangled then /etc/master.passwd can be used as a reference copy.

sudo su root, why not sudo -i?

There's an i flag for sudo that replaces sudo su?

meena: su needs -l too

and 'su root' is equivalent to 'su'

Hi, I have successfully installed FreeBSD 13.2 w/ GELI and ZFS. And I know how to make custom partition layout. But when I tried to add authentication feature to the geli device, I failed to boot with message "Can't find /boot/zfsloader | /boot/loader | /boot/kernel/kernel". What should I do for a bootable geli device with authentication?

After whatever drm/mesa thing got updated and "fixed" my blender issue, I'm getting hard resets when using blender. Can anyone advise on how to figure out which component is even at fault? Far as I can tell this could be caused anywhere between drm-kmod, gpu firmware, mesa stuff, vulkan/opengl, x11 and blender and I seemingly can't even get a single line of output telling me what went wrong…

kern.powercycle_on_panic and kern.poweroff_on_panic are 0, kern.panic_reboot_wait_time is 15 (should all be default) so as far as I can tell a "normal" kernel panic shouldn't even instantly reset the machine, so I'm pretty much at a complete loss on how to figure this one out.

neither current nor previous dmesg and messages in /var/log seem to contain even a single line about what happened either… i just open blender, work with a negligible amount of vertices and poof system's gone. variable amount of time, not doing anything specific. sometimes it just happens while rotating the camera. only commonality is that it happens while blender is redrawing its viewport

(moving vertices or camera, essentially)

i have the normal dumpdev="AUTO" in my rc.conf, but /var/crash only has a "minfree" file whose only content is "2048", nothing else.

yeah.. well i get an error with u-boot BUG /drivers/usb/host/xhci-rb.c:548/abort_id()

that was a goo dmess up right there

well look at that, it boots.. i am cluseless

still have no user root error but i am able to boot into the environment

well it loks like i have this toor account but now i can not login.. did learn of this toor account.. had no clue about it before

if you can use sudo in your regular login, you can 'sudo su toor'

sudo: unknown user root

sudo:error initializing audit plugin sudoers_audit

huh, unexpected, that

ahhh

sudo is broken too

try su toor

i am presented password, i try my root password

and says no bueno (sorry)

blank password does the same

maybe i should try making a password for it

well that did not work.. it says passwd: permission denied (which makes sense, since not doing it as root)

don't think you can, if you're not a superuser account (root or toor)

yeah

are you able to boot to single user mode?

well at least i got my uncommited code off it and it was able to "boot" looks like this rainy sunday is doing what you suggestd

i will have to give that another try, gotta boot the machine on workbench

brb, going to set that up right now, to see what i can get

I think it is better to boot off a FreeBSD liveCD and inspect from there

yeah you're right

other alternatives - su -l toor or su -m toor

either way, need to move machines over and should have a bootable usb thumbdrive also.. learning from mistakes

just tried su -l toor and su -m toor and they wont work if no passwd is set for toor

i did notice they fixed esolution issues on raspberry pi install

now i am not blinded by the txt being huge on this monitor

err i did the no boot menu option

i do not like that option to get rid of boot menu and do -1

beastie_disable="YES" is bad

Hi

It seems it was not motherboard issue

here is the errors in dmesg (I copied everything under network initialisation since that was when I plugged in the USB sound card that works on FreeBSD laptop)

f451: toor usually has a password of *

It gets io_error and can't set USB buss

so you cannot su to that account

If it perhaps CPU issue and not MotherBoard? Or do I need some firmware or something?

these three lines are interesting from the dmesg:

xhci0: Resetting controller

usbd_req_re_enumerate: addr=5, set address failed! (USB_ERR_TIMEOUT, ignored)

usbd_setup_device_desc: getting device descriptor at addr 5 failed, USB_ERR_IOERROR

either your USB device is faulty or the USB controller/port

or the cable

I tried USB on other laptop and PC, no issue, with same cable

Just changed motherboard because I thought it was motherboard issue

Googling a bit and maybe bios setting, or perhaps it is cable. I will go get the original cable it came with (much shorter)

Is there any possibility it is CPU related?

I tried using the usb cable it came with, same issue. I am off to bios

Ok so it says the usb is self powered, does it mean it thinks my USB soundcard has it's own powersource or that it is powering the USB?

My sound card gets power from USB

is it using usb1, 2 or 3?

It is a USB2 device but I connect to USB3

do you have any usb2 ports to try it in?

Yes. Last time no issue iirc, let me try again

I try another USB 3 on back side now

Ok, no keyboard or mouse issues at least

Let's see if sound breaks or not (Last time I got it to work on other motherboard it had a lot of crackling, which I suspect is from not enough juice being fed)

Ok it seems to work properly now. Perhaps front panel USB 3 doesn't give juice other than the 100mA or whatever USB drives run at

Previous motherBoard had a lot of crackling

I guess previous motherboard was shit. Although this time I change a BIOS setting about XHCI devices

Could have been that

Aw well, thanks for the tip, I had given up on back USB being different from front since prior MoBo had same result

a

are there any known bugs in the updated ix driver ?

I just enabled it and it says it detected the device is in firmware recovery mode and it cuts off the connection

but the device is most definitely not in firmware recovery mode and it runs the latest firmware from Intel, 3.60

I also run the 3.3.35 package version from Intel

i have options GDB set (13.2-release) but debug.kdb.available does not contain any entries. my intention is to use bhyve -G. wondering if there is anything i'm missingthat i could be doing wrong that is preventing the GDB backend from being available

I did a freebsd-update on 13.2-RELEASE, where does the system keep the original sshd_config file ?

mns: Looks like /var/db/freebsd-update/files. Maybe check something like /var/db/freebsd-update/*-rollback/INDEX-OLD for the gzip filename to 'zcat'.

Or, if you happen to be using ZFS, freebsd-update probably created a snapshot.

It appears I can't record from my mic

Mumble doesn't hear anything

psychonate: thanks. I'll take a look

I didn't realize there is a binary patch now as well for that recent zfs issue: freebsd.org/security/advisories/FreeBSD-EN-23:16.openzfs.asc

when i log in, my system tells me i'm on releng/14.0-n265380, but the revision that fixes the ZFS issue is releng/14.0-n265384. does that mean this fix wasn't included in 14.0-RELEASEp1?

scoobybejesus: it probably just means that uname wasn't updated

or whatever the heck uname is reading

is that a new thing with 14.0 ? I don't get a -nXXXXXXXX after the release nuber in 13.2-RELEASE-p4

scoobybejesus, Is your system compiled from source or freebsd-update binary? I am guessing it is compiled from source and that is the source version.

Also is that the base userland version or kernel version?

actually, i am all binary.. no source

i actually only am able to see that version by logging in via ssh and reading the login message. maybe there is another way. i haven't looked deeper.

I have not yet updated to 14 so I don't know but... The base userland version is hard coded into the script. "less /bin/freebsd-version" and see USERLAND_VERSION there.

Oh! The official way is "freebsd-version -kru" for the full set of three versions. The installed kernel, the running kernel, which may be different, and the base userland.

By inspection of the script we see the base userland version is "compiled" into that script and reported directly from there.

i guess it's in uname -a as well

The installed kernel is found by using "what -qs /boot/kernel/kernel" which on my 13 system reports "FreeBSD 13.2-RELEASE-p4 GENERIC".

And the running kernel is reported by asking the kernel itself with "sysctl kern.osrelease" which reports this on my system "kern.osrelease: 13.2-RELEASE-p4".

freebsd-version -kru prints as expected, the third adding -p1

The kernel and the base userland get updated separately and are frequently not the same if updates were released for one but nothing changed in the other.

I'm not sure what all 'uname' checks, but I think the 14.0 p1 binary update didn't touch /boot/kernel/kernel, so that may be why only the userland has the p1.

I am sure uname reports the kernel's struct uname data which will be the same as the "sysctl kern.osrelease" data.

but it did update /boot/kernel/zfs.ko and /lib/libzpool.so and some other files, so that's a good sign

Title: FreeBSD Errata Notice FreeBSD-EN-23:16.openzfs

psychonate: do you get -p1 for 'freebsd-version -kr'

mns: No, only for 'freebsd-version -u', but I think that is expected. The binary patch didn't need to update /boot/kernel/kernel. It updated a zfs lib and the zfs module, I believe.

damn I am speachless. I am not a player, and watching now video of NightCity from CyberPunk 2077. Amazing. Unbelievable.

anyway. Is FreeBSD a hobby for you? Or what? How can you describe your relation ;)

and how about programming knowledle. Do you code? Are you proffesional or hibbyst?

iam just curious

wikan: i'm an admin, turned coder, I do so semi professionally for FreeBSD's benefit. but it's more of an expensive hobby than a job at this point with how much slippage i have piled up

what is coding for you meena personaly? Is it some king of art, technical ability, sience?

JFYI, it's art. See en.wikipedia.org/wiki/The_Art_of_Computer_Programming

Title: The Art of Computer Programming - Wikipedia

wikan: all the above, and more

all the above or maybe there is one in particular?