On 3/23/16 5:43 , Guenther Alka wrote:
> For me, Solarish, Supermicro mainboards, Avago HBAs and Intel nics
> are nearly a guarantee for a trouble free Illumos setup.
> 
> Now there is a new 1151/ X11-SS generation of mainboards available, that
> might be the the number one solution for an entry level to midrange
> barebone or virtualized storage solution with IPMI, up to 64GB ECC RAM,
> CPUs like the dualcore G4400 up from 50 Euro to fast quadcore Xeons
> both with ECC and vt-d support.
> 
> There are models with 1G, 10G and with or without Avago/LSI 3008 HBA.
> I now got one of the first X11SSH-CTF boards in Germany with HBA and 10G.
> http://www.supermicro.com/products/motherboard/Xeon/C236_C232/X11SSH-CTF.cfm
> 
> 
> I was able to install ESXi 6u2 but when I tried a barebone install of
> OmniOS
> 151016 or 151017, after booting from USB, I got a:
> 
> - Probing for device nodes
>> Console login, service cannot run,
> enter user for system maintenance mode:
> 
> 
> As I know, that the X550 10G nic is not yet supported, I disabled the nic
> with a jumper setting and then the Avago 3008 HBA but with same result.
> Boot order was set to legacy (EFI disabled)
> 
> 
> A quick check with Oracle Solaris 11.3 gives the same result
> Has anyone seen the same problem and has an idea how to solve?

Hi Guenther,

Whenever these kinds of problems crop up, I know they can be quite
frustrating, but with your help, we should be able to take it root
cause, though it may be a bit involved. The very first thing I'd do is
make sure I'm on an IPMI serial console and not some VGA console.

The very first thing I do here is I boot to milestone=none, to help get
a sense of where things might be going on. Usually these hangs are
driver related, but sometimes I've seen things get gummed up for other
reasons (though I don't have much OmniOS installer experience).

The next thing I usually do is use kmdb. I add -v to the boot options
and see what the last thing I see is. Note, that this last thing can be
a little misleading, as it's often the last successsful thing that we see.

Usually this means it's going to end up being a more manual process as I
add -kd to the kernel command line in grub, which will drop you into
kmdb. From kmdb I set the value of moddebug to 0x80000001 by running:

> moddebug/W 0x80000001

This will cause us to trap into the kernel debugger every time we load a
module. So what I do here is I use this to see what the last thing I see
is. Once I have run that, I end up typing a lot of ':c' and trying to
figure out what the last one I see is. Then once I have that, I'll go
back and not using moddebug, set a breakpoint in that module's init or
attach function and see what the last instruction I see is by stepping
over things and seeing where exactly it's hanging.

Does this help?

Robert


-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175430-2e6923be
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21175430&id_secret=21175430-6a77cda4
Powered by Listbox: http://www.listbox.com

Reply via email to