[illumos-gate] 16645 Core DDI function device/DMA attributes could be const -- Robert Mustacchi <rm⊙fo>

hi

does illumos implement 802.11ac protocol?

Hello. Can someone quickly guide me through expanding my root "/" ZFS filesystem on OI to the/my new maximum disk size? The system is a KVM VM, I expanded the disk image from 80GB to 160GB. I already had a look at the "format" utility from illumos. I don't know how grow/extend the partition size inside the 'format' utility. I searched with Google, but all results I found are incomplete. Thank you in

advance.

As I found out, root@openindiana ~ # diskinfo alread sees the new disk size: - c6t0d0 Virtio Block Device 160.00 GiB

Inside the partition> print mode, I have two partitions, root and backup. I don't know what to do here, if I am allowed to simply alter their size (both? as they are both the same size?). How can I get/receive the maximum size of the disk, while altering the selected partition? There is also a boot partition. I assume I leave this partition the way it is? Thanks.

show zpool list from the system

NAME SIZE ALLOC FREE CKPOINT EXPANDSZ FRAG CAP DEDUP HEALTH ALTROOT

rpool 79,5G 61,5G 18,0G - - 31% 77% 1.00x ONLINE -

try zpool online -e rpool c6t0d0

cannot expand c6t0d0: no such device in pool

zpool status rpool

and specify the device

root and backup you say? it means you first need to expand the Solaris2 partition with fdisk (format -> select the disk -> enter fdisk), then, go to partition menu in format and set new size for zfs slice by entering slice number, then use the same starting point and for size use $

NAME STATE READ WRITE CKSUM

rpool ONLINE 0 0 0

c6t0d0s0 ONLINE 0 0 0

huh, s0, how did you install the system?

then enter partition command to store new partition table on disk and after that you can expand your pool

denk: I installed the system with the GUI installation ISO.

denk: It is legacy BIOS installation.

Aedil: paste.dilos.org/?3aeff9f1bab93c50#7…HUSpMH8BWlUGYHCRGxvlJdHNQaQuVENbxA=

denk: Shall I use zpool online -e rpool c6t0d0s0?

[illumos-gate] 16666 loader: reading from large dosfs fails -- Toomas Soome <tsoome⊙mc>

Aedil: do what tsoome said

denk: Ok.

I can not remember from top of my head if we did implement zpool online -e support with GPT partition table (it may be done), but MBR+Solaris2 partitioning is manual work to fix the partition tables first.

and in general, I'd suggest to use GPT (whole disk setup) if possible.

tsoome: Do I need to "delete" and then "create" the whole partition with format> fdisk ? There is now grow/expand menu item here.

um, i think it was the only way, yes

100% space as you probably do not need to support other OS'es there:)

add a new device to the vm, create a pool, then zfs send+zfs recv, install a boot loader, try to boot from a new disk

tsoome: Thanks, I did the fdisk part. Now it says 100%. I will continue with your second suggestion.

Solaris2 partition is your fdisk partition, inside that you have VTOC partition scheme, so you need to fix that one too.

and since there may be custom fdisk partitioning + custom vtoc slices, automatic partition update from zpool is rather annoying thing to implement with such setup.

I made the changes, but it won't boot afterwards. I am now in the rescue disk. How can I re-apply the boot loader? Thanks.

I booted into the OI GUI ISO.

zpool import -N rpool, then bootadm install-bootloader

bootadm install-bootloader -P rpool

always double check the check the manual first;)

tsoome: Thank you!

tsoome: I will try now!

May it be that it now does the zpool expand, before showing the bootloader? I have 100% usage of one core, but it still displays nothing. I will wait.

zpool list, compare the SIZE column

30 minutes ago you had it 79,5G

It was 89GB when I re-installed the bootloader. There is still load on one CPU core.

30 minutes passed by so far.

How long is the ZFS autoexpand from 80GB to 160GB to take (in time)?

It should be essentially instantaneous

I was able to ZFS import rpool from the OI ISO disquette. So the file system most probably is not lost.

One core is loaded and it says Booting from Hard Disk, no further error message.

45 minutes passed by, I am not very firm into ZFS.

What does the system do right now? It says Booting from Hard Disk… no further message. It must be ZFS related.

One core is loaded.

A good hour passed by.

When I do zpool import rpool from the rescue disk, the /rootfs mount is empty, only /boot in there.

The /home mount is still with all data.

It still says 89GB size.

How can I recover from this?

zpool list now says 160 GB :-)

ALLOC 61.5 GB and Free 98.0 GB.

I can still see my beadm mountpoints.

Hi, I'm currently messing with a Windows Domain Controller. I'm trying to get rid of the outdated NTLM authentication. But I've noticed, that the OmniOS test file server which is joined to the domain uses NTLMv2 to authenticate. Can illumos/OmniOS use Kerberos instead?

Can someone help me? My OpenIndiana VM does not boot anymore. It booots into a black screen, one core 100% loaded, no bootloader appearing, after modifying the Solaris2 partition table. Thanks.

The also had set zpool set autoexpand=on.

system

tsoome_: denk: I was able to fix it! Thank you, guys for your help. I had to run # bootadm install-bootloader -f -M -P rpool from the installation disquette! Thank you so much. Only after adding the "-M" switch to install into the BIOS MBR, it worked again! I am so lucky now. :-)=

The filesystem also was expanded to 160G size. Everything is excellent now! I hope I can remember the steps I learned today.

The manpage of bootadm and docs.oracle.com/cd/E36784_01/html/E36801/glzhs.html helped me to fix the issue.

Aedil: you need to break (and to recover) the system 10 (or more) times to understand every step

also we can help to break it too :)

unixery yes, we can use kerberos

Aedil you are welcome.

tsoome_: How can I do it? I've joined the domain via "smbadm join -y -u myadmin mydomain.local"

I've found nothing about kerberos in smbadm man page

Is the current OI illumos branched as SunOS 5.11.6? I have seen once something like SunOS 5.11.6 for illumos. The original Oracle Solaris stays at SunOS 5.11.4, whis is why I ask. Most probably, in the future, newer release numbers will or must be used. Thanks for any insights.

root@openindiana ~ # pkg update -v – Error encountered while retrieving data from: [Errno 5] I/O error

Is the IPS pkg manager download from the OI illumos mirror broken right now?

Solaris uname -rs stays at "SunOS 5.11" - no .4 there, the .4 is only in uname -v output which is 11.4.... without the "5."

I've never seen anyone list a 5.11.x output in any uname for Solaris or illumos

alanc: Ok, thank you.

Today I have read on heise.de that NVIDIA wants to focus on open-source drivers.

wants or is getting pressured because of anti-trust concerns surrounding AI in the EU?

but at least they're going to be dual GPL and MIT licensed

Indeed!

It will be interesting to see how useful it is without closed source userland binaries etc

for which?

Not terribly as there's a bunch of firmware.

That needs to get uploaded to the GPU, IIRC, which took on more management parts of it with the move to their custom risc-v core.

if enough of it to please katago can port, that'd please at least me

Also, this is something they've been doing for a few years now, so not anything new.

I'd imagine, like, libGL is a whole other kettle of fish

alanc: I don't think anyone user-facing changes the 5.11, since it would confuse autoconf, etc, to such a large degree.

I think appliance type stuff changes the whole thing

I was never able to convince people we should have gone to 6.0 for S11 since it was a major release you can't upgrade to from S10 - I was going to try again for S13 since it was going to drop 32-bit compatibility, but that plan died when they made me roll back 5.12 to 5.11

but yeah, I expect everything in actual use is 5.11 now

5.11 4eva knuckle tats