basic xend problems

[Jukka Salmi <j+nbsd%2010.salmi.ch@localhost>]

Hi,

I'm having troubles setting up Xen 3.3 on a NetBSD/amd64 5.1_RC1 system
with xen{kernel,tools}33 from pkgsrc.

Dom0 boots fine.  Then I start xend and xenbackendd using the rc.d
scripts shipping with the xentools33 package.  Now I'm able to create a
domU (`xm create -u ...'), use it and destroy it.  So far so good.  But
while stopping xenbackendd now succeeds, stopping xend fails:

  $ sudo /etc/rc.d/xenbackendd onestop
  Stopping xenbackendd.
  Waiting for PIDS: 463.
  $ sudo /etc/rc.d/xend onestop
  Stopping xend.
  Waiting for PIDS: 958, 958, 958, 958, 958, 958, 958, 958, 958, 958, 958^C
  $ ps -ax | grep xen
    97 ?     Il   0:00.01 xenconsoled
   958 ?     Il   0:00.45 /usr/pkg/bin/python2.6 /usr/pkg/sbin/xend start
   963 ?     I    0:00.12 xenstored --pid-file /var/run/xenstore.pid
  $ sudo kill 958 963 97
  $ ps -ax | grep xen
   958 ?     Il   0:00.46 /usr/pkg/bin/python2.6 /usr/pkg/sbin/xend start
  $ sudo kill -9 958
  $ ps -ax | grep xen

When xend received the SIGTERM, it logged

  [2010-04-30 13:55:10 958] DEBUG (XMLRPCServer:235) XMLRPCServer.cleanup()

but apparently did not die until it was sent a SIGKILL.

Trying to start xend again now fails with the following being logged:

  [2010-04-30 13:59:06 965] INFO (SrvDaemon:331) Xend Daemon started
  [2010-04-30 13:59:06 965] INFO (SrvDaemon:335) Xend changeset: unavailable.
  [2010-04-30 13:59:06 965] ERROR (SrvDaemon:347) Exception starting xend 
(Shared object "libexpat.so.1" not found)
  Traceback (most recent call last):
    File "/usr/pkg/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.py", 
line 339, in run
      servers = SrvServer.create()
    [...]
    File "/usr/pkg/lib/python2.6/site-packages/_xmlplus/parsers/expat.py", line 
4, in <module>
      from pyexpat import *
  ImportError: Shared object "libexpat.so.1" not found
  [2010-04-30 13:59:06 1022] INFO (SrvDaemon:219) Xend exited with status 1.

The system in question does not have a libexpat.so since I did not extract the
xbase set.  Do I need the X sets to use Xen?  However, removing the directory
/var/run/xend/state seems to fix this problem and xend starts again.

But now, when trying to create my domU again (`xm create -c ...'), I don't see
any console output anymore, but the domU (nbsd) seems to be created:

  $ sudo xm list
  Name                                        ID   Mem VCPUs      State   
Time(s)
  Domain-0                                     0  1023     1     r-----     21.2
  nbsd                                         2   128     1     --p---      0.0

One hunderd seconds later the domU is automatically (?!) destroyed, and
xend's log shows:

  [2010-04-30 14:06:57 1562] INFO (XendDomainInfo:1907) createDevice: vbd : 
[...]
  [...]
  [2010-04-30 14:06:57 1562] DEBUG (XendDomainInfo:1450) 
XendDomainInfo.handleShutdownWatch
  [2010-04-30 14:06:57 1562] DEBUG (DevController:166) Waiting for devices vif.
  [2010-04-30 14:06:57 1562] DEBUG (DevController:171) Waiting for 0.
  [2010-04-30 14:06:57 1562] DEBUG (DevController:656) hotplugStatusCallback 
/local/domain/0/backend/vif/2/0/hotplug-status.
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:2409) 
XendDomainInfo.destroy: domid=2
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:1946) Destroying device model
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:1953) Releasing devices
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:1966) Removing vif/0
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:910) 
XendDomainInfo.destroyDevice: deviceClass = vif, device = vif/0
  [2010-04-30 14:08:37 1562] ERROR (XendDomainInfo:1973) Device release failed: 
nbsd; vif; vif/0
  Traceback (most recent call last):
    File "/usr/pkg/lib/python2.6/site-packages/xen/xend/XendDomainInfo.py", 
line 1967, in _releaseDevices
      self.destroyDevice(true_devclass, dev, False);
    File "/usr/pkg/lib/python2.6/site-packages/xen/xend/XendDomainInfo.py", 
line 924, in destroyDevice
      rc = self.getDeviceController(deviceClass).destroyDevice(devid, force)
    File 
"/usr/pkg/lib/python2.6/site-packages/xen/xend/server/DevController.py", line 
255, in destroyDevice
      self.writeBackend(dev, 'online', "0")
    File 
"/usr/pkg/lib/python2.6/site-packages/xen/xend/server/DevController.py", line 
501, in writeBackend
      raise VmError("Device %s not connected" % devid)
  VmError: Device 0 not connected
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:1966) Removing console/0
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:910) 
XendDomainInfo.destroyDevice: deviceClass = console, device = console/0
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:1966) Removing vbd/1
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:910) 
XendDomainInfo.destroyDevice: deviceClass = vbd, device = vbd/1
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:1951) No device model
  [2010-04-30 14:08:37 1562] DEBUG (XendDomainInfo:1953) Releasing devices

The only workaround I found so far to make the domU work again is to
start xend after the dom0 has been rebooted...  Any hints?

Help is appreciated.


TIA, Jukka

-- 
This email fills a much-needed gap in the archives.