Re: Failure to boot, Debian squeeze with 4.0.1 hypervisor, timer problems?

[Andy Smith <andy@strugglers.net>]

Hi Jan,

Thanks for getting back to me.

On Mon, Jul 30, 2012 at 08:58:03AM +0100, Jan Beulich wrote:
> >>> On 27.07.12 at 18:15, Andy Smith <andy@strugglers.net> wrote:
> > I upgraded it to Debian squeeze, with xen-hypervisor-4.0-amd64
> > (4.0.1-4) / linux-image-2.6.32-5-xen-amd64 (2.6.32-45) and now it
> > does not complete a boot of the dom0 kernel.
> 
> The minimum would be to check against 4.0.3 (ideally tip of
> 4.0-testing), to exclude a problem already fixed even on that
> branch.

I checked out http://xenbits.xen.org/hg/xen-4.0-testing.hg and
compiled the hypervisor. It seems to boot okay, though I have not
yet done any exhaustive stress tests.

"xm info" now says:

    host                   : sol
    release                : 2.6.32-5-xen-amd64
    version                : #1 SMP Sun May 6 08:57:29 UTC 2012
    machine                : x86_64
    nr_cpus                : 4
    nr_nodes               : 1
    cores_per_socket       : 4
    threads_per_core       : 1
    cpu_mhz                : 2133
    hw_caps                : bfebfbff:28100800:00000000:00001b40:009ce3bd:00000000:00000001:00000000
    virt_caps              : hvm
    total_memory           : 24567
    free_memory            : 23280
    node_to_cpu            : node0:0-3
    node_to_memory         : node0:23280
    node_to_dma32_mem      : node0:2996
    max_node_id            : 0
    xen_major              : 4
    xen_minor              : 0
    xen_extra              : .4-rc4-pre
    xen_caps               : xen-3.0-x86_64 xen-3.0-x86_32p hvm-3.0-x86_32 hvm-3.0-x86_32p hvm-3.0-x86_64 
    xen_scheduler          : credit
    xen_pagesize           : 4096
    platform_params        : virt_start=0xffff800000000000
    xen_changeset          : Mon Jul 30 13:39:47 2012 +0100 21607:6d7ae840463c
    xen_commandline        : placeholder dom0_mem=1024M loglvl=all guest_loglvl=all com1=9600,8n1 console=com1,vga serial_tx_buffer=64k
    cc_compiler            : gcc version 4.4.5 (Debian 4.4.5-8) 
    cc_compile_by          : andy
    cc_compile_domain      : strugglers.net
    cc_compile_date        : Mon Aug  6 02:40:47 UTC 2012
    xend_config_format     : 4

(a keen observer will note that my last set of logs showed 96G RAM,
whereas it now has only 24G. I had to take those 6x16G DIMMs out and
use them for something else in the meantime, and I replaced them
with 6x4G ones. I did confirm that the problem was still evident
before carrying on.)

Is there any point at this stage in trying to find out which commit
appears to have fixed the problem I was seeing? If so, any pointers
on the best ways to do that would be appreciated as I'm not that
familiar with hg (or git).

I am not ready to use 4.1 in production at the moment so I have to
stick with 4.0.x. I'm assuming that if I wish to continue using
4.0.x on this host I would be better off compiling the released
4.0.3 and using that (if it works for me), as opposed to the tip of
4.0-testing that I have here?

Thanks,
Andy