[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] OpenSolaris zombie domains

  • To: xen-devel@xxxxxxxxxxxxxxxxxxx, xen-discuss@xxxxxxxxxxxxxxx
  • From: Robert Nelson <robertn@xxxxxxxxxxxxxxx>
  • Date: Sat, 10 Nov 2007 19:26:57 -0800
  • Delivery-date: Sat, 10 Nov 2007 19:30:14 -0800
  • Domainkey-signature: a=rsa-sha1; s=the-nelsons.org; d=the-nelsons.org; c=simple; q=dns; h=message-id:from; b=ZCCo5GhoCaxTaMXRk6R5mf/EimYPOgAxJ2GbJdDxXnrBbasufnL6WS633WoT IAB0UlKFul/J1YHCeGacjXPsFaD6X1+FA1NVhjKMSiKBdTlanQzPkUasL +X6VL1dBfh/joSyG1xMLlMhr5mwTrFaT6E6hOwRmrCKkzgGmyUeJ0Y=;
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>

I'm running xen 3.1 with Fedora 8 as dom0.  When I shut down an OpenSolaris domU it leaves a zombie domain with state = shutdown:1 and dying:1.  There is one page left allocated.

Here is the result of the "q" command on xen console:
(XEN) General information for domain 3:
(XEN)     refcnt=1 nr_pages=1 xenheap_pages=0 dirty_cpus={}
(XEN)     handle=a30786a8-e959-2c0c-6cbc-d0f4683d4213 vm_assist=00000004
(XEN) Rangesets belonging to domain 3:
(XEN)     Interrupts { }
(XEN)     I/O Memory { }
(XEN)     I/O Ports  { }
(XEN) Memory pages belonging to domain 3:
(XEN)     DomPage 0000000118cfd000: mfn=0000000000118cfd, caf=00000001, taf=0000000080000001
(XEN) VCPU information and callbacks for domain 3:
(XEN)     VCPU0: CPU0 [has=F] flags=0 upcall_pend = 01, upcall_mask = 00 dirty_cpus={} cpu_affinity={0-31}
(XEN)     100 Hz periodic timer (period 10 ms)
(XEN)     Notifying guest (virq 1, port 3, stat -1/0/-1)
>From my limited understanding of xen internals it appears that this is a level 4 page table.  I believe this is a xen bug since it also happens if I destroy the domain rather than shutting down gracefully.

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.