[oi-dev] hipster 2021-10 failed to boot after pkg update

[Toomas Soome]

> On 28. May 2022, at 16:29, Klaus Elsbernd <Klaus.Elsbernd at posteo.de> wrote:
> 
> Urgent help needed :-(
> 
> a server (supermicro/xeon) has been installed and upgraded to 2021-10. It was last updated on 2021:12:12. Today I updated the packages with
> 
> pkg update
> 
> Now the system failed to boot into multi-user/single-user mode with the following services not started:
> 
> Enter user name tor system maintenance (control-d to bypass): root
> Enter root password (control-d to bypass): single-user privilege assigned to root on /dev/conso1e.
> Entering System Maintenance Mode
> May 28 14:47:56 su: ' su root' succeeded for root on /dev/console
> The illumos Project illumos-b0d58672df May 2022 root at serv-1100:/root# svcs -x svc:/system/metainit:default (SVM initialization)
> State: maintenance since Sat May 28 14:46:50 2022
> Reason: Completes a dependency cycle.
> See: http://i1lumos.org/msg/SMF-8000-HP
> See: metainit(lM)
> Impact: This service is not running.
> root at serv-1100:/root# beadm list
> BE    Active    Mountpoint    Space    Policy Created
> openindiana-2020:06:11    -    -    62.03M    static 2020-06-11 16:53
> openindiana-2020:08:10  -    -    11.97M    static 2020-08-10 16:44
> openindiana-2020:12:29  -    -    46.41M    static 2020-12-29 13:54
> openindiana-2021:01:29  -    -    22.00M    static 2021-01-29 12:37
> openindiana-2021:05:06  -    -    11.88M    static 2021-05-06 19:01
> openindiana-2021:05:06-backup-l  -    -    128K    static 2021-05-14 15:17
> openindiana-2021:06:05  -    -    1.906    static 2021-06-05 12:21
> openindiana-2021:06:05-1  -    -    16.68M    static 2021-06-05 20:08
> openindiana-2021:12:12  -    -    14.90M    static 2021-12-12 17:40
> 
> open!ndiana-2022:05:28    NR    /  43.356    static 2022-05-28 11:01
> 
> The machine is booting OpenIndiana Hipster 2021.10 version illumos-b0d58672df 64-bit. Some notable errors
> 
> acpinex: cpu at 0, cpudrv0
> /fw/sb at 0/socket at 0/cpu at 0 (cpudrv0) online
> pseudo-device: cld0
> cld0 is Zpseudo/cld at 0
> sorry, variable 'ip_squeue_worker_wait' is not defined in the 'ip' module
> PCI Express-device: pciBBBB,3cB2 at l. pcieb0
> 
> One message appeared on the console, indicating an error with a NVIDIA module. But there is no NVIDIA graphic card installed:
> 
> /kernel/drv/amd64l/nvidia.modeset: undefined symbol 'nvidia_get_rm_ops'
> WARNING: mod_load: cannot load module 'nvidia.modeset'
> WARNING: nvidia_modeset: unable to resolve dependency, module 'nvidia' not found
> 
> This is the second machine I lost. The only solution was then to boot the previous boot environment 2021:12:12.
> 
> Any ideas?
> 
> Klaus
> 

If you do not have nvidia card, then this message is annoying, but harmless - because there is nocard, nvidia driver will not get loaded and module nvidia.modeset also does not load.

this other message may also be hint - "sorry, variable 'ip_squeue_worker_wait' is not defined in the 'ip’ module”. Do you have this variable set in /etc/system? you can try to boot -a and when it does prompt for /etc/system, enter /dev/null

rgds,
toomas