[OpenIndiana-discuss] BTX halted

[Nikola M]

On 05/ 6/17 01:12 PM, Jim Klimov wrote:
> On May 6, 2017 10:33:31 AM GMT+03:00, Nikola M <minikola at gmail.com> wrote:
>
>> I see 201704 'release notes' are out there but I think that 20150502 is
>> testing and not the snapshot that is considered 201704 snapshot, still
>> and there is the time to recompile illumos for it, remove power button
>> security issue and test.
Here should be 20170502 test snapshot, not 2015.. of course. ;)
> By the way, as another checkpoint - my VirtualBoxed installation on the laptop (still with GRUB) was pkg-updated early last weekend (and worked after reboot), and then on 3rd morning - and failed to boot. I ended up with a blank screen where X11 should've started. 

Check if you maybe have too many Boot Environments for GRUB to handle
and also, there is a mention that VirualBoox guest add-ons need
reinstallation.

> Logging in on the serial console, I saw that some networking services were disabled that I think should have been running, and a couple were in maintenance. I enabled and cleared those, rebooted - but still had no GUI (although no unexpected not-online services now).

I have failure to restart lightdm service in recent few months, but it
fails to start just after being in long session for some time (fails to
start after logout), but restarts if just re-logging.
Don't know if it is related, but I got some lines in log that point to
some Nvidia driver failure, could be connected with something similar in
VirtualBox?
Mine is able to start after full restart, even it holds and stay there
frozen on restart and shutdow, in does restart/shutdown.

[ 55936.338] (EE) NVIDIA(GPU-0): EVO Push buffer channel allocation failed
[ 55936.339] (EE) NVIDIA(GPU-0): Failed to allocate EVO core DMA push buffer
[ 56025.101] (EE) NVIDIA(0): Failing initialization of X screen 0
in application-graphical-login-lightdm:default.log
and several
[ 55930.626]     compiled for 1.18.4, module version = 1.0.0
in Xorg.0.log
(could be that nvidia driver is not set for newer environment and not
targeted for OI)
> We were running off for a trip, so I had no time to write down the details or investigate, and I'll next see that laptop next week probably. My guess is that in the couple of days the Xorg changes had landed and perhaps just a reinstall (recompile) of Vbox guest additions might fix things; but without digging I can't rule out something more sinister yet ;)
>
> So just anecdotal experience from a single experiment, at the moment. 
I suppose so.
Btw, my booting problem is reproducible only on ACPI v1 machines and
only on hardware, but I suppose many people still have those nice older
64bit machines working, so ISOs ned to be changed not to loose many
people's new installations.

When venting and discussing in getneral already,
main issue is that people should not hasitate to report their bugs as
soon as they appear and do that 'en masse'. Only way anything to be
fixed is to report it ASAP, not waiting. For example, during /hipster
repository updates, from one OI snapshot to another, one can surf
through osnet-incorporation and userland-incorporation incorporations,
to find out _when_ certain bug first appeared.
But after the OI snapshot, those previous states are gone/deleted, to
free disk space, to speed up IPS/pkg(5) operations, to lower RAM usage
for pkg, so not any bug reporting is too soon. :)