Re: [Xen-devel] [xen-unstable test] 4607: regressions - FAIL

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Keir Fraser writes ("Re: [Xen-devel] [xen-unstable test] 4607: regressions - 
FAIL"):
> On 11/01/2011 12:22, "Ian Jackson" <Ian.Jackson@xxxxxxxxxxxxx> wrote:
> > This ought to get us pushes more often.  It would still be nice to get
> > rid of some of these race bugs :-).
> 
> Where are the races, do you think?

Known races:

 * Sometimes, after running xl destroy, you can find the following
   things left over in xenstore:
      /local/domain/0/device-model/16 = ""   (n0)
      /local/domain/0/device-model/16/command = "shutdown"   (n0)
   This is a race in libxl domain destruction, probably.
   Causes failures in
      test-*-*-xl*  leak-check

 * The XCP 2.6.27 kernel has a bug where it signals a block device
   creation through the udev machinery to userland before the device
   is actually readable.  This results in the initramfs failing to
   find its root filesystem.  Normally this happens during a domU
   startup but I have seen it once or twice in a dom0.
   Causes failures in
     test-*-xcpkern-*         xen-boot
     test-*-xcpkern-{xl,pv}*  guest-start

 * Mysterious occasional guest startup failures.  PV guests for some
   reason don't produce a guest console logfile.  Happens with xend
   and xl and PV and HVM.  Could conceivably be due to some occasional
   network infrastructure problems in the test environment.

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel