hello all, a simple question, for the repos: github.com/joyent/linux-live.git it requires btrfs / lvm related packages. my question is, since it is contains zfs in the live image, is it possible to not contain the packages for the image?

and 2nd question, is it possible to export iscsi target in smartos global zone without lost the information when the GZ reboot?

is the new freebsd-loader based image EFI only by any chance

ok recent platform images crash on boot with grub

i can't remember for the life of me what i did to fix it cause the last drive i used failed

page_upsub bad argument: pp 0, *ppp f000eef3`f000ee00

*vpsub

every single boot

unable to go into rescue mode either

i turned off SR-IOV and the unused IOMMU and that seemed to help

since illumos has _never_ supported the AMD IOMMU

hmmm... i have a service to allow modifications to the system in the global zone, namely to have an ansible user in this case, and it used to work just fine and ansible could login over ssh with pubkeys.. but now it fails, and i can't even su to ansible, it says: "su: No shell" ... but i see the ansible user in /etc/passwd and the shell is correct, the user exists in /etc/shadow, what else could this be?

i don't know if something changed during an upgrade or if i've changed something that broke it in the mean time, but what else could it be when the user is in /etc/passwd and the shell binary has the correct (not modified) permissions?

<despair86> "since illumos has _never_..." <- does that mean bHyve does not work on AMD CPUs in Illumos?

bhyve works fine, but no PCI passthrough

in fact

bhyve is the *only* option for VMs on AMD64

KVM was always EM64T

on illumos

also can someone paste the result of mount | grep shadow ? my service runs on boot and i'm wondering if i inverted the order of the arguments to the mount command there; i don't think it's related to the above (shell is in /etc/passwd) but int he interest of correctness..

# usermod -s `which bash` dafuq; su dafuq

su: No shell

ok, n evermind. i found the error

@despair86 I believe PCI Passthrough works with AMD now

pmooney: isn't that so?

don't believe so

at least not on illumos

the existing IOMMU driver is (afaik) hopelessly busted, and the one from bhyve hasn't been ported

ahhh, good to know

it is admittedly not a high prio right now

speaking for myself