still can't start domain

still can't start domain

[James Harper]

[-- Attachment #1: Type: text/plain, Size: 14300 bytes --]

is there still a problem starting domains or is it just me? I get this error:

xen1:~# xm create -f /etc/xen/mail2
Traceback (most recent call last):
  File "/usr/sbin/xm", line 6, in ?
    main.main(sys.argv)
  File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 507, in main
    xm.main(args)
  File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 71, in main
    self.main_call(args)
  File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 88, in main_call
    p.main(args[1:])
  File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 162, in main
    create.main(args)
  File "/usr/lib/python2.3/site-packages/xen/xm/create.py", line 400, in main
    (d, c) = make_domain(opts, config)
  File "/usr/lib/python2.3/site-packages/xen/xm/create.py", line 361, in make_domain
    dominfo = server.xend_domain_create(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendClient.py", line 219, in xend_domain_create
    {'op'      : 'create',
  File "/usr/lib/python2.3/site-packages/xen/xend/XendClient.py", line 150, in xend_call
    return xend_request(url, "POST", data)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendClient.py", line 121, in xend_request
    raise RuntimeError(resp.reason)
RuntimeError: Error creating domain

and this output from xend:

config: (vm (name 'mail2 vm 1') (memory '128') (cpu '1') (image (linux (kernel /boot/vmlinuz-2.4.26-xenU) (ip ::::vm1:eth0:off) (root '/dev/sda8 ro') (args ' 4 VMID=1 '))) (device (vbd (uname phy:sda8) (dev sda8) (mode w))) (device (vif (mac FE:FD:00:00:03:00) (bridge br0))) (device (vif (mac FE:FD:00:00:03:01) (bridge br1))))
vm_create>
create_domain> linux /boot/vmlinuz-2.4.26-xenU
create-domain> init_domain...
init_domain> 128 mail2 vm 1 1
create_domain> dom= 3
build_domain> linux 3 /boot/vmlinuz-2.4.26-xenU  ip=::::vm1:eth0:off root=/dev/sda8 ro  4 VMID=1
VIRTUAL MEMORY ARRANGEMENT:
 Loaded kernel: c0000000->c0144b98
 Init. ramdisk: c0145000->c0145000
 Phys-Mach map: c0145000->c0165000
 Page tables:   c0165000->c0167000
 Start info:    c0167000->c0168000
 Boot stack:    c0168000->c0169000
 TOTAL:         c0000000->c0400000
 ENTRY ADDRESS: c0000000
vm_create<
EVENT> xend.console.create [13, 3, 9603]
respond_be_create>
>create_devices
blkif_dev_create> 3 2056 w {'device': 2048, 'type': '83', 'start_sector': 86204853L, 'nr_sectors': 5879727L}
<create_devices
_vm_configure1> made devices...
_vm_configure1<
recv_be_vbd_create> <xen.xend.server.blkif.BlkifControllerFactory instance at 0x40224ccc>
brctl addif br0 vif3.0
vm_dev_vif> created <xen.xend.server.netif.NetDev instance at 0x4078a14c>
recv_be_vbd_grow> <xen.xend.server.blkif.BlkifControllerFactory instance at 0x40224ccc>
cleanup> 3
release_devices> 3
release_vifs> 3
brctl delif br0 vif3.0
Traceback (most recent call last):
  File "/etc/xen/xend/vifctl", line 149, in ?
    VifControl().main(sys.argv)
  File "/etc/xen/xend/vifctl", line 49, in main
    meth(args[1:])
  File "/etc/xen/xend/vifctl", line 141, in vifctl_down
    if params['ip']:
KeyError: 'ip'
brctl delif nbe-br vif3.1
device vif3.1 is not a slave of nbe-br
Traceback (most recent call last):
  File "/etc/xen/xend/vifctl", line 149, in ?
    VifControl().main(sys.argv)
  File "/etc/xen/xend/vifctl", line 49, in main
    meth(args[1:])
  File "/etc/xen/xend/vifctl", line 141, in vifctl_down
    if params['ip']:
KeyError: 'ip'
release_vbds> 3
>BlkifController>send_be_disconnect> dom= 3
op_create> Deferred Exception creating domain: [Failure instance: Traceback: <type 'tuple'>, (<twisted.python.failure.Failure twisted.internet.defer.TimeoutError>, 2)
]
Unhandled error in Deferred:
(debug:  C: Deferred was created:
 C:  File "/usr/sbin/xend", line 45, in ?
 C:    sys.exit(main())
 C:  File "/usr/sbin/xend", line 33, in main
 C:    return daemon.start()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 521, in start
 C:    self.run()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 633, in run
 C:    reactor.run()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/default.py", line 126, in run
 C:    self.mainLoop()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/default.py", line 137, in mainLoop
 C:    self.doIteration(t)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/pollreactor.py", line 145, in doPoll
 C:    log.callWithLogger(selectable, _drdw, selectable, fd, event, POLLIN, POLLOUT, log)
 C:  File "/usr/lib/python2.3/site-packages/twisted/python/log.py", line 65, in callWithLogger
 C:    callWithContext({"system": lp}, func, *args, **kw)
 C:  File "/usr/lib/python2.3/site-packages/twisted/python/log.py", line 52, in callWithContext
 C:    return context.call({ILogContext: newCtx}, func, *args, **kw)
 C:  File "/usr/lib/python2.3/site-packages/twisted/python/context.py", line 43, in callWithContext
 C:    return func(*args,**kw)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/pollreactor.py", line 160, in _doReadOrWrite
 C:    why = selectable.doRead()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 253, in doRead
 C:    self.protocol.notificationReceived(notification)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 177, in notificationReceived
 C:    channel.notificationReceived()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/channel.py", line 132, in notificationReceived
 C:    self.handleNotification()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/channel.py", line 298, in handleNotification
 C:    work += self.handleResponses()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/channel.py", line 335, in handleResponses
 C:    self.responseReceived(msg)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/channel.py", line 343, in responseReceived
 C:    dev.responseReceived(msg, ty, subty)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/controller.py", line 92, in responseReceived
 C:    if self.callResponders(msg):
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/controller.py", line 140, in callResponders
 C:    resp.responseReceived(msg)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/controller.py", line 31, in responseReceived
 C:    self.deferred.callback(msg)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 252, in callback
 C:    self._startRunCallbacks(result)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 313, in _startRunCallbacks
 C:    self._runCallbacks()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
 C:    self.result = callback(self.result, *args, **kw)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 124, in respond_be_create
 C:    d.callback(blkif)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 252, in callback
 C:    self._startRunCallbacks(result)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 313, in _startRunCallbacks
 C:    self._runCallbacks()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
 C:    self.result = callback(self.result, *args, **kw)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 314, in _vm_configure1
 C:    d = vm.create_devices()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 661, in create_devices
 C:    v = dev_handler(self, dev, dev_index)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 806, in vm_dev_vif
 C:    defer = make_vif(vm.dom, vif, vmac, vm.recreate)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 156, in make_vif
 C:    d = xend.netif_dev_create(dom, vif, vmac, recreate=recreate)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 745, in netif_dev_create
 C:    d = ctrl.attachDevice(vif, vmac, recreate=recreate)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/netif.py", line 270, in attachDevice
 C:    d = defer.Deferred()
 I: First Invoker was:
 I:  File "/usr/sbin/xend", line 45, in ?
 I:    sys.exit(main())
 I:  File "/usr/sbin/xend", line 33, in main
 I:    return daemon.start()
 I:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 521, in start
 I:    self.run()
 I:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 633, in run
 I:    reactor.run()
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/default.py", line 126, in run
 I:    self.mainLoop()
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/default.py", line 134, in mainLoop
 I:    self.runUntilCurrent()
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/base.py", line 423, in runUntilCurrent
 I:    call.func(*call.args, **call.kw)
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 385, in <lambda>
 I:    lambda: self.called or timeoutFunc(self, *args, **kw))
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 154, in timeout
 I:    deferred.errback(failure.Failure(TimeoutError("Callback timed out")))
)
Failure: twisted.internet.defer.TimeoutError: Callback timed out
(netif (dom '3'))
>BlkifController>send_be_destroy> dom= 3
Unhandled error in Deferred:
(debug:  C: Deferred was created:
 C:  File "/usr/sbin/xend", line 45, in ?
 C:    sys.exit(main())
 C:  File "/usr/sbin/xend", line 33, in main
 C:    return daemon.start()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 521, in start
 C:    self.run()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 633, in run
 C:    reactor.run()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/default.py", line 126, in run
 C:    self.mainLoop()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/default.py", line 134, in mainLoop
 C:    self.runUntilCurrent()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/base.py", line 423, in runUntilCurrent
 C:    call.func(*call.args, **call.kw)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 385, in <lambda>
 C:    lambda: self.called or timeoutFunc(self, *args, **kw))
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 154, in timeout
 C:    deferred.errback(failure.Failure(TimeoutError("Callback timed out")))
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 274, in errback
 C:    self._startRunCallbacks(fail)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 313, in _startRunCallbacks
 C:    self._runCallbacks()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
 C:    self.result = callback(self.result, *args, **kw)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 491, in _cbDeferred
 C:    self.errback(failure.Failure((result, index)))
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 274, in errback
 C:    self._startRunCallbacks(fail)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 313, in _startRunCallbacks
 C:    self._runCallbacks()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
 C:    self.result = callback(self.result, *args, **kw)
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 294, in _continue
 C:    self.unpause()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 290, in unpause
 C:    self._runCallbacks()
 C:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
 C:    self.result = callback(self.result, *args, **kw)
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 438, in cberr
 C:    self.destroy()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 512, in destroy
 C:    self.cleanup()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 531, in cleanup
 C:    self.release_devices()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 542, in release_devices
 C:    self.release_vifs()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 553, in release_vifs
 C:    ctrl.destroy()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/netif.py", line 256, in destroy
 C:    self.destroyDevices()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/netif.py", line 260, in destroyDevices
 C:    dev.destroy()
 C:  File "/usr/lib/python2.3/site-packages/xen/xend/server/netif.py", line 170, in destroy
 C:    d = defer.Deferred()
 I: First Invoker was:
 I:  File "/usr/sbin/xend", line 45, in ?
 I:    sys.exit(main())
 I:  File "/usr/sbin/xend", line 33, in main
 I:    return daemon.start()
 I:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 521, in start
 I:    self.run()
 I:  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 633, in run
 I:    reactor.run()
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/default.py", line 126, in run
 I:    self.mainLoop()
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/default.py", line 134, in mainLoop
 I:    self.runUntilCurrent()
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/base.py", line 423, in runUntilCurrent
 I:    call.func(*call.args, **call.kw)
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 385, in <lambda>
 I:    lambda: self.called or timeoutFunc(self, *args, **kw))
 I:  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 154, in timeout
 I:    deferred.errback(failure.Failure(TimeoutError("Callback timed out")))
)
Failure: twisted.internet.defer.TimeoutError: Callback timed out

James

[-- Attachment #2: Type: text/html, Size: 18214 bytes --]

Re: still can't start domain

[ron minnich]

whew!

these python things are a bear to debug when they go wrong ... 

you could try strace on this thing to see what's failing. That's a hard 
backtrace for me to figure out.

ron



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Ian Pratt]

> whew!
> 
> these python things are a bear to debug when they go wrong ... 

Yep, the asynchronous nature of Twisted does make it tough to
figure out where the real problem is. The current approach is
just to print every possible back trace rather than trying to be
smart. 

At a glance, the error appears to be here:

Traceback (most recent call last):
  File "/etc/xen/xend/vifctl", line 149, in ?
    VifControl().main(sys.argv)
  File "/etc/xen/xend/vifctl", line 49, in main
    meth(args[1:])
  File "/etc/xen/xend/vifctl", line 141, in vifctl_down
    if params['ip']:
KeyError: 'ip'

It looks like we're not passing in an ip= parameter into
vifctl. Since vifctl is about to die and be replaced with a
rather simpler/shorter shell script, I'll pass on this one.

Best,
Ian


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Rune Johan Andresen]

Yes, I have the same problem. I get this error msg when I type "start 
xend"

File "/usr/sbin/xend", line 23, in ?
	from xen.xend.server import SrvDaemon
      File 
"/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 
19
	in ?
	from twisted.internet import pollreactor
ImportError: No module named twisted.internet

(after installing all the tools)

Any ideas?



On Jul 8, 2004, at 8:05 AM, Ian Pratt wrote:

>> whew!
>>
>> these python things are a bear to debug when they go wrong ...
>
> Yep, the asynchronous nature of Twisted does make it tough to
> figure out where the real problem is. The current approach is
> just to print every possible back trace rather than trying to be
> smart.
>
> At a glance, the error appears to be here:
>
> Traceback (most recent call last):
>   File "/etc/xen/xend/vifctl", line 149, in ?
>     VifControl().main(sys.argv)
>   File "/etc/xen/xend/vifctl", line 49, in main
>     meth(args[1:])
>   File "/etc/xen/xend/vifctl", line 141, in vifctl_down
>     if params['ip']:
> KeyError: 'ip'
>
> It looks like we're not passing in an ip= parameter into
> vifctl. Since vifctl is about to die and be replaced with a
> rather simpler/shorter shell script, I'll pass on this one.
>
> Best,
> Ian
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
> digital self defense, top technical experts, no vendor pitches,
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Tom Wilkie]

It looks like you need to install twisted, or have it set up properly.

You can get the package from here:

http://twisted.sourceforge.net/Twisted_NoDocs-1.3.0.tar.gz

Follow the instructions on installing it in there, they're slightly 
weird.

Tom

On 8 Jul 2004, at 09:57, Rune Johan Andresen wrote:

>
> Yes, I have the same problem. I get this error msg when I type "start 
> xend"
>
> File "/usr/sbin/xend", line 23, in ?
> 	from xen.xend.server import SrvDaemon
>      File 
> "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDaemon.py", line 
> 19
> 	in ?
> 	from twisted.internet import pollreactor
> ImportError: No module named twisted.internet
>
> (after installing all the tools)
>
> Any ideas?
>
>
>
> On Jul 8, 2004, at 8:05 AM, Ian Pratt wrote:
>
>>> whew!
>>>
>>> these python things are a bear to debug when they go wrong ...
>>
>> Yep, the asynchronous nature of Twisted does make it tough to
>> figure out where the real problem is. The current approach is
>> just to print every possible back trace rather than trying to be
>> smart.
>>
>> At a glance, the error appears to be here:
>>
>> Traceback (most recent call last):
>>   File "/etc/xen/xend/vifctl", line 149, in ?
>>     VifControl().main(sys.argv)
>>   File "/etc/xen/xend/vifctl", line 49, in main
>>     meth(args[1:])
>>   File "/etc/xen/xend/vifctl", line 141, in vifctl_down
>>     if params['ip']:
>> KeyError: 'ip'
>>
>> It looks like we're not passing in an ip= parameter into
>> vifctl. Since vifctl is about to die and be replaced with a
>> rather simpler/shorter shell script, I'll pass on this one.
>>
>> Best,
>> Ian
>>
>>
>> -------------------------------------------------------
>> This SF.Net email sponsored by Black Hat Briefings & Training.
>> Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
>> digital self defense, top technical experts, no vendor pitches,
>> unmatched networking opportunities. Visit www.blackhat.com
>> _______________________________________________
>> Xen-devel mailing list
>> Xen-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/xen-devel
>
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital 
> self defense, top technical experts, no vendor pitches, unmatched 
> networking opportunities. Visit www.blackhat.com
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel
>



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Ian Pratt]

> It looks like you need to install twisted, or have it set up properly.
> 
> You can get the package from here:
> 
> http://twisted.sourceforge.net/Twisted_NoDocs-1.3.0.tar.gz
> 
> Follow the instructions on installing it in there, they're slightly 
> weird.

'make install-twisted' may be helpful, depending on your setup.

Ian


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[David Becker]

On Thu, Jul 08, 2004 at 10:41:50AM +1000, James Harper wrote:
" is there still a problem starting domains or is it just me? I get this error:

I couldn't start a domain with Tuesday's pull.  With Wednesday's pull,
xm seems to start a domain (based on execution time reported by xm list) but
I can't get any console output or network response from the created domain.

The previous tool set had much better debuggability even if it was less tidy.
What is the draw for using twisted?



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Mike Wray]

Ian Pratt wrote:

>>whew!
>>
>>these python things are a bear to debug when they go wrong ... 
> 
> 
> Yep, the asynchronous nature of Twisted does make it tough to
> figure out where the real problem is. The current approach is
> just to print every possible back trace rather than trying to be
> smart. 
> 
> At a glance, the error appears to be here:
> 
> Traceback (most recent call last):
>   File "/etc/xen/xend/vifctl", line 149, in ?
>     VifControl().main(sys.argv)
>   File "/etc/xen/xend/vifctl", line 49, in main
>     meth(args[1:])
>   File "/etc/xen/xend/vifctl", line 141, in vifctl_down
>     if params['ip']:
> KeyError: 'ip'
> 
> It looks like we're not passing in an ip= parameter into
> vifctl. Since vifctl is about to die and be replaced with a
> rather simpler/shorter shell script, I'll pass on this one.
> 

The problem is that vifctl is installed as a config file - so
it's not installed if it's already there. And the parameter
name has been changed to 'ipaddr'. If you delete the old vifctl and
replace it with the uptodate one it'll be ok.

Mike


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[ron minnich]

On Thu, 8 Jul 2004, David Becker wrote:

> The previous tool set had much better debuggability even if it was less tidy.
> What is the draw for using twisted?

this worries me too. It reminds me of the Perl async sockets library that
Intermezzo used to use. It was essentially impossible to debug. It was 
eventually replaced by C code. 

Of all the tools Xen has set up so far I most prefer the 'small set of
simple C programs' approach. I was most happy with that.

ron



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Mark Williamson]

> The previous tool set had much better debuggability even if it was less tidy.
> What is the draw for using twisted?

Twisted makes it nicer to write servers based on nonblocking IO.  I found I 
quite liked it, once I got my head around it.  In includes generic support for 
doing nonblocking IO, handling protocols, scheduling deferred work, etc.  
There's also some higher-level stuff but I don't know about that.

The control tools have historically been rather simple and some have got a bit 
ugly over time.  They worked pretty well but we really needed this new 
functionality to help us to support enterprise-level use.  Unfortunately, 
doing things "properly" does come at the price of some increased complexity.  
The extra work now should pay off in stability, ease of use and quicker 
development in the future.

Hopefully we can add features to make debugging less of an issue as this new 
code matures.  Any suggestions that could improve debug-ability,etc are (as 
always) encouraged.

Cheers,
Mark



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Avery Pennarun]

On Thu, Jul 08, 2004 at 04:15:09PM +0100, Mark Williamson wrote:

> Hopefully we can add features to make debugging less of an issue as this new 
> code matures.  Any suggestions that could improve debug-ability,etc are (as 
> always) encouraged.

Clearer error messages than "Error: Cannot start domain" would probably be a
good start.  Anyone could have figured out in short order that the domain
couldn't be started.  What's much more important is *why* the domain
couldn't be started.

In general, error messages should always say why an operation failed, not
just that the operation failed.  I find that this works wonders in my own
programs.  In C programs, for example, always print out 'errno' in any error
message where it could possibly be relevant.  (Of course, to avoid
confusion, definitely don't print out errno where it's not valid anymore :))
The errno equivalent in perl is $!.  In python I guess you have to use
exceptions or something.

Have fun,

Avery


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Ian Pratt]

> On Thu, Jul 08, 2004 at 04:15:09PM +0100, Mark Williamson wrote:
> 
> > Hopefully we can add features to make debugging less of an issue as this new 
> > code matures.  Any suggestions that could improve debug-ability,etc are (as 
> > always) encouraged.
> 
> Clearer error messages than "Error: Cannot start domain" would probably be a
> good start.  Anyone could have figured out in short order that the domain
> couldn't be started.  What's much more important is *why* the domain
> couldn't be started.
> 
> In general, error messages should always say why an operation failed, not
> just that the operation failed.  I find that this works wonders in my own
> programs.  In C programs, for example, always print out 'errno' in any error
> message where it could possibly be relevant.  (Of course, to avoid
> confusion, definitely don't print out errno where it's not valid anymore :))
> The errno equivalent in perl is $!.  In python I guess you have to use
> exceptions or something.

There's actually serious effort being made to provide good error
messages. Xen always returns a useful error code on failed
operations, which the Xc library turns into python
exceptions. xend picks up the exception, which is a little
complicated given the asyncronous nature of xend. It then needs
to communicate the text of the exception message back to the
client ('xm') so that it can print a useful message for the user.

At the moment, there's reams of debuging and traceing messages
enabled, so the useful error message gets completely swamped in
the middle of a bunch of other messages useful only to the xend
developer (Mike). This will obviously be quietened down before we
move toward a 2.0 release candidate.

I think the new xend is the right direction, but its integration
has been way, way, more painful than we ever anticipated or hoped
for.  Mike has been working very hard to try and sort out the
various issues, and I'm sure we'll get to a stable, feature-full
xend soon. Please be patient for a little while longer...

Cheers,
Ian



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Keir Fraser]

> Clearer error messages than "Error: Cannot start domain" would probably be a
> good start.  Anyone could have figured out in short order that the domain
> couldn't be started.  What's much more important is *why* the domain
> couldn't be started.

This is absolutely true.

> In general, error messages should always say why an operation failed, not
> just that the operation failed.  I find that this works wonders in my own
> programs.  In C programs, for example, always print out 'errno' in any error
> message where it could possibly be relevant.  (Of course, to avoid
> confusion, definitely don't print out errno where it's not valid anymore :))
> The errno equivalent in perl is $!.  In python I guess you have to use
> exceptions or something.

The problem with the new toolset is that some of the context for
building a useful error message at the client program is separated
from the client in both space (its in a server, behind a socket) and
in time (the failure may be due to some random asynchronous request
made ages ago). A tedious problem but certainly surmountable -- this
area definitely needs attention.

The crux of the problem is that we need a domain controller capable of
juggling multiple outstanding operations at the same time. In general
you do that by using multiple threads or by decoupling requests from
their eventual responses. Both can have their drawbacks, and
asynchronous models in particular need great care to avoid runaway
complexity.

My preferred model would be lightweight language-level threads, but I
don't know if anything suitable exists for Python. What I've seen of
Twisted so far hasn't recommended it to me. :-)

 -- Keir


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[ron minnich]

On Thu, 8 Jul 2004, Keir Fraser wrote:

> The problem with the new toolset is that some of the context for
> building a useful error message at the client program is separated
> from the client in both space (its in a server, behind a socket) and
> in time (the failure may be due to some random asynchronous request
> made ages ago). A tedious problem but certainly surmountable -- this
> area definitely needs attention.

yes, this is how it played out on intermezzo, and I don't think in the 
limit it was solvable. Thus this highly async model in Perl was dropped. 
It just became un-debuggable.

ron



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Brian Wolfe]

[-- Attachment #1: Type: text/plain, Size: 1133 bytes --]

Hmmm, this is sort of what i'm doing with xenctld in C. vhosts, nodes, and master nodes all have states. Actions are taken based on the state and input is parsed and compared to the states to determine the next step. It operates similar to how xend and xc_dom_control interact, a main daemon that does the work and a client that issues and reports to the end user. The communications all happen in XML (still working on the DTD) so it should be easy for 3rd party stuff to hook into it.

My first test program just to play around is the C version.

The next attempt at the core features will be done in python just to prove that it works and be able to rapidly change features, etc.

Alas, we are a LONG way from it being usefull though. Just been doing the design specs and the network protocol specs. I intend to make some money from the extended features of xenctld and release a subset of capabilities as a GPL version eventually. So for now, don't hold your breath. I just wanted everyone to know that there are a couple of us that are concentrating on making a feature rish set of tools that will eventually be available. :)



[-- Attachment #2: Type: text/plain, Size: 84 bytes --]

On Thu, 08 Jul 2004 17:15:28 +0100
 Keir Fraser <Keir.Fraser@cl.cam.ac.uk> said...


[-- Attachment #3: Type: text/plain, Size: 2142 bytes --]

> Clearer error messages than "Error: Cannot start domain" would probably be a
> good start.  Anyone could have figured out in short order that the domain
> couldn't be started.  What's much more important is *why* the domain
> couldn't be started.

This is absolutely true.

> In general, error messages should always say why an operation failed, not
> just that the operation failed.  I find that this works wonders in my own
> programs.  In C programs, for example, always print out 'errno' in any error
> message where it could possibly be relevant.  (Of course, to avoid
> confusion, definitely don't print out errno where it's not valid anymore :))
> The errno equivalent in perl is $!.  In python I guess you have to use
> exceptions or something.

The problem with the new toolset is that some of the context for
building a useful error message at the client program is separated
from the client in both space (its in a server, behind a socket) and
in time (the failure may be due to some random asynchronous request
made ages ago). A tedious problem but certainly surmountable -- this
area definitely needs attention.

The crux of the problem is that we need a domain controller capable of
juggling multiple outstanding operations at the same time. In general
you do that by using multiple threads or by decoupling requests from
their eventual responses. Both can have their drawbacks, and
asynchronous models in particular need great care to avoid runaway
complexity.

My preferred model would be lightweight language-level threads, but I
don't know if anything suitable exists for Python. What I've seen of
Twisted so far hasn't recommended it to me. :-)

 -- Keir


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

Re: still can't start domain

[Mark Williamson]

> I think the new xend is the right direction, but its integration
> has been way, way, more painful than we ever anticipated or hoped
> for.

I think part of the reason is that there are lots of changes under the hood so 
that things are done really properly.  The new tools form a much better 
platform to build on and should facilitate the implementation of new features 
(e.g. cluster control, etc).

Like Ian said, things should quieten down in the nearish future.  And by 2.0 
it will really rock :-)

Mark



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Rune Johan Andresen]

Hi, after install all the necessary packages I also have problems to make
everything function:

xend start gives:


 brctl addbr nbe-br
device nbe-br already exists; can't create bridge with the same name
brctl sethello nbe-br 0
brctl setfd nbe-br 0
brctl stp nbe-br off
Interface not found: eth0
initial_refresh>
db dom= ['domain', ['id', '0'], ['name', 'Domain-0'], ['memory', '119']]
xc dom= {'cpu_time': 637937478041L, 'name': 'Domain-0', 'dying': 0, 'dom':
0, 'maxmem_kb': 131072, 'paused': 0, 'crashed': 0, 'running': 1,
'shutdown_reason': 0, 'shutdown': 0, 'mem_kb': 121856, 'cpu': 0,
'blocked': 0}
dom= 0 config= ['domain', ['id', '0'], ['name', 'Domain-0'], ['memory',
'119']]
dom= 0 new
XendDomain>refresh>
XendDomain>initial_refresh> doms:
dom domain id=0 name=Domain-0 memory=119
EVENT> xend.start 0
XendDomain>reap>
XendDomain>reap<


COMMENTS:

(Comment to error msg: eth0 should be found! ifconfig shows that it is set:

eth0      Link encap:Ethernet  HWaddr 00:07:E9:DF:E7:D4
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:220697 errors:0 dropped:0 overruns:0 frame:0
          TX packets:3110 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:61640932 (58.7 MiB)  TX bytes:215917 (210.8 KiB)
          Interrupt:11 Base address:0xd000
)



xm create gives:

 File "/usr/sbin/xm", line 6, in ?
    main.main(sys.argv)
  File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 473, in main
    xm.main(args)
  File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 71, in main
    self.main_call(args)
  File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 88, in
main_call
    p.main(args[1:])
  File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 162, in main
    create.main(args)
  File "/usr/lib/python2.3/site-packages/xen/xm/create.py", line 395, in main
    opts.load_defaults()
  File "/usr/lib/python2.3/site-packages/xen/xm/opts.py", line 334, in
load_defaults
    self.load(p, help)
  File "/usr/lib/python2.3/site-packages/xen/xm/opts.py", line 359, in load
    execfile(defaults, globals, locals)
  File "/etc/xen/xmdefaults", line 21, in ?
    xm_vars.check()
  File "/usr/lib/python2.3/site-packages/xen/xm/help.py", line 45, in check
    v.doCheck(self.env)
  File "/usr/lib/python2.3/site-packages/xen/xm/help.py", line 75, in doCheck
    env[self.name] = self.check(self.name, env.get(self.name))
  File "/etc/xen/xmdefaults", line 11, in vmid_check
    val = int(val)
TypeError: int() argument must be a string or a number

Hope someone have time to see if they find anything interessting in these
tracebacks. :-)

Rune J.A





-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Rune Johan Andresen]

Hi, after install all the necessary packages I also have problems to 
make
everything function:

xend start gives:


  brctl addbr nbe-br
device nbe-br already exists; can't create bridge with the same name
brctl sethello nbe-br 0
brctl setfd nbe-br 0
brctl stp nbe-br off
Interface not found: eth0
initial_refresh>
db dom= ['domain', ['id', '0'], ['name', 'Domain-0'], ['memory', '119']]
xc dom= {'cpu_time': 637937478041L, 'name': 'Domain-0', 'dying': 0, 
'dom':
0, 'maxmem_kb': 131072, 'paused': 0, 'crashed': 0, 'running': 1,
'shutdown_reason': 0, 'shutdown': 0, 'mem_kb': 121856, 'cpu': 0,
'blocked': 0}
dom= 0 config= ['domain', ['id', '0'], ['name', 'Domain-0'], ['memory',
'119']]
dom= 0 new
XendDomain>refresh>
XendDomain>initial_refresh> doms:
dom domain id=0 name=Domain-0 memory=119
EVENT> xend.start 0
XendDomain>reap>
XendDomain>reap<


COMMENTS:

(Comment to error msg: eth0 should be found! ifconfig shows that it is 
set:

eth0      Link encap:Ethernet  HWaddr 00:07:E9:DF:E7:D4
           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
           RX packets:220697 errors:0 dropped:0 overruns:0 frame:0
           TX packets:3110 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:1000
           RX bytes:61640932 (58.7 MiB)  TX bytes:215917 (210.8 KiB)
           Interrupt:11 Base address:0xd000
)



xm create gives:

  File "/usr/sbin/xm", line 6, in ?
     main.main(sys.argv)
   File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 473, in 
main
     xm.main(args)
   File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 71, in 
main
     self.main_call(args)
   File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 88, in
main_call
     p.main(args[1:])
   File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 162, in 
main
     create.main(args)
   File "/usr/lib/python2.3/site-packages/xen/xm/create.py", line 395, 
in main
     opts.load_defaults()
   File "/usr/lib/python2.3/site-packages/xen/xm/opts.py", line 334, in
load_defaults
     self.load(p, help)
   File "/usr/lib/python2.3/site-packages/xen/xm/opts.py", line 359, in 
load
     execfile(defaults, globals, locals)
   File "/etc/xen/xmdefaults", line 21, in ?
     xm_vars.check()
   File "/usr/lib/python2.3/site-packages/xen/xm/help.py", line 45, in 
check
     v.doCheck(self.env)
   File "/usr/lib/python2.3/site-packages/xen/xm/help.py", line 75, in 
doCheck
     env[self.name] = self.check(self.name, env.get(self.name))
   File "/etc/xen/xmdefaults", line 11, in vmid_check
     val = int(val)
TypeError: int() argument must be a string or a number

Hope someone have time to see if they find anything interessting in 
these
tracebacks. :-)

Rune J.A


On Jul 8, 2004, at 9:03 PM, Mark Williamson wrote:

>> I think the new xend is the right direction, but its integration
>> has been way, way, more painful than we ever anticipated or hoped
>> for.
>
> I think part of the reason is that there are lots of changes under the 
> hood so
> that things are done really properly.  The new tools form a much better
> platform to build on and should facilitate the implementation of new 
> features
> (e.g. cluster control, etc).
>
> Like Ian said, things should quieten down in the nearish future.  And 
> by 2.0
> it will really rock :-)
>
> Mark
>
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
> digital self defense, top technical experts, no vendor pitches,
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Keir Fraser]

Problem is that the bridge-setup code expects the Ethernet interface
to have an IP address. If you really mean to not have an IP address on
eth0, then we can fix the code so that it will work for you.

 -- Keir

> Hi, after install all the necessary packages I also have problems to 
> make
> everything function:
> 
> xend start gives:
> 
> 
>   brctl addbr nbe-br
> device nbe-br already exists; can't create bridge with the same name
> brctl sethello nbe-br 0
> brctl setfd nbe-br 0
> brctl stp nbe-br off
> Interface not found: eth0
> initial_refresh>
> db dom= ['domain', ['id', '0'], ['name', 'Domain-0'], ['memory', '119']]
> xc dom= {'cpu_time': 637937478041L, 'name': 'Domain-0', 'dying': 0, 
> 'dom':
> 0, 'maxmem_kb': 131072, 'paused': 0, 'crashed': 0, 'running': 1,
> 'shutdown_reason': 0, 'shutdown': 0, 'mem_kb': 121856, 'cpu': 0,
> 'blocked': 0}
> dom= 0 config= ['domain', ['id', '0'], ['name', 'Domain-0'], ['memory',
> '119']]
> dom= 0 new
> XendDomain>refresh>
> XendDomain>initial_refresh> doms:
> dom domain id=0 name=Domain-0 memory=119
> EVENT> xend.start 0
> XendDomain>reap>
> XendDomain>reap<
> 
> 
> COMMENTS:
> 
> (Comment to error msg: eth0 should be found! ifconfig shows that it is 
> set:
> 
> eth0      Link encap:Ethernet  HWaddr 00:07:E9:DF:E7:D4
>            UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>            RX packets:220697 errors:0 dropped:0 overruns:0 frame:0
>            TX packets:3110 errors:0 dropped:0 overruns:0 carrier:0
>            collisions:0 txqueuelen:1000
>            RX bytes:61640932 (58.7 MiB)  TX bytes:215917 (210.8 KiB)
>            Interrupt:11 Base address:0xd000
> )
> 
> 
> 
> xm create gives:
> 
>   File "/usr/sbin/xm", line 6, in ?
>      main.main(sys.argv)
>    File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 473, in 
> main
>      xm.main(args)
>    File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 71, in 
> main
>      self.main_call(args)
>    File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 88, in
> main_call
>      p.main(args[1:])
>    File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 162, in 
> main
>      create.main(args)
>    File "/usr/lib/python2.3/site-packages/xen/xm/create.py", line 395, 
> in main
>      opts.load_defaults()
>    File "/usr/lib/python2.3/site-packages/xen/xm/opts.py", line 334, in
> load_defaults
>      self.load(p, help)
>    File "/usr/lib/python2.3/site-packages/xen/xm/opts.py", line 359, in 
> load
>      execfile(defaults, globals, locals)
>    File "/etc/xen/xmdefaults", line 21, in ?
>      xm_vars.check()
>    File "/usr/lib/python2.3/site-packages/xen/xm/help.py", line 45, in 
> check
>      v.doCheck(self.env)
>    File "/usr/lib/python2.3/site-packages/xen/xm/help.py", line 75, in 
> doCheck
>      env[self.name] = self.check(self.name, env.get(self.name))
>    File "/etc/xen/xmdefaults", line 11, in vmid_check
>      val = int(val)
> TypeError: int() argument must be a string or a number
> 
> Hope someone have time to see if they find anything interessting in 
> these
> tracebacks. :-)
> 
> Rune J.A
> 
> 
> On Jul 8, 2004, at 9:03 PM, Mark Williamson wrote:
> 
> >> I think the new xend is the right direction, but its integration
> >> has been way, way, more painful than we ever anticipated or hoped
> >> for.
> >
> > I think part of the reason is that there are lots of changes under the 
> > hood so
> > that things are done really properly.  The new tools form a much better
> > platform to build on and should facilitate the implementation of new 
> > features
> > (e.g. cluster control, etc).
> >
> > Like Ian said, things should quieten down in the nearish future.  And 
> > by 2.0
> > it will really rock :-)
> >
> > Mark
> >
> >
> >
> > -------------------------------------------------------
> > This SF.Net email sponsored by Black Hat Briefings & Training.
> > Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
> > digital self defense, top technical experts, no vendor pitches,
> > unmatched networking opportunities. Visit www.blackhat.com
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/xen-devel
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
> digital self defense, top technical experts, no vendor pitches, 
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Niraj Tolia]

On Fri, 09 Jul 2004 11:56:31 +0100, Keir Fraser
<keir.fraser@cl.cam.ac.uk> wrote:
> 
> Problem is that the bridge-setup code expects the Ethernet interface
> to have an IP address. If you really mean to not have an IP address on
> eth0, then we can fix the code so that it will work for you.
> 
>  -- Keir
> 
> 

Keir,

I am getting the same error. However, its the bridging code itself
that setups nbe-br and causes eth0 not to have an IP address. So any
xend stop followed by a xend start will give the error message.

Still haven't figured out why xm create is failing though. The
traceback occurs because check_vmid is trying to convert the value
'None' into a int. Need to look at why the wrong value is being passed
through but someone else who knows the code better than me would have
better luck.

Niraj


[snip]


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Rune Johan Andresen]

Is this a common problem for the unstable ver. or is this an exception 
for me and Niraj? ref: "xm create" - >
(I am able to start xend without error msgs now) Maybe I should use the 
1.2 ver for benchmarking, or does the
unstable perform better?

Rune



> xm create gives:
>
>   File "/usr/sbin/xm", line 6, in ?
>      main.main(sys.argv)
>    File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 473, in
> main
>      xm.main(args)
>    File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 71, in
> main
>      self.main_call(args)
>    File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 88, in
> main_call
>      p.main(args[1:])
>    File "/usr/lib/python2.3/site-packages/xen/xm/main.py", line 162, in
> main
>      create.main(args)
>    File "/usr/lib/python2.3/site-packages/xen/xm/create.py", line 395,
> in main
>      opts.load_defaults()
>    File "/usr/lib/python2.3/site-packages/xen/xm/opts.py", line 334, in
> load_defaults
>      self.load(p, help)
>    File "/usr/lib/python2.3/site-packages/xen/xm/opts.py", line 359, in
> load
>      execfile(defaults, globals, locals)
>    File "/etc/xen/xmdefaults", line 21, in ?
>      xm_vars.check()
>    File "/usr/lib/python2.3/site-packages/xen/xm/help.py", line 45, in
> check
>      v.doCheck(self.env)
>    File "/usr/lib/python2.3/site-packages/xen/xm/help.py", line 75, in
> doCheck
>      env[self.name] = self.check(self.name, env.get(self.name))
>    File "/etc/xen/xmdefaults", line 11, in vmid_check
>      val = int(val)
> TypeError: int() argument must be a string or a number




On Jul 9, 2004, at 1:18 PM, Niraj Tolia wrote:

> On Fri, 09 Jul 2004 11:56:31 +0100, Keir Fraser
> <keir.fraser@cl.cam.ac.uk> wrote:
>>
>> Problem is that the bridge-setup code expects the Ethernet interface
>> to have an IP address. If you really mean to not have an IP address on
>> eth0, then we can fix the code so that it will work for you.
>>
>>  -- Keir
>>
>>
>
> Keir,
>
> I am getting the same error. However, its the bridging code itself
> that setups nbe-br and causes eth0 not to have an IP address. So any
> xend stop followed by a xend start will give the error message.
>
> Still haven't figured out why xm create is failing though. The
> traceback occurs because check_vmid is trying to convert the value
> 'None' into a int. Need to look at why the wrong value is being passed
> through but someone else who knows the code better than me would have
> better luck.
>
> Niraj
>
>
> [snip]
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
> digital self defense, top technical experts, no vendor pitches,
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Niraj Tolia]

Rune,

So the way you pass the vmid on the command line changed and I was
slightly confused by the new documentation (from xm help create).

Just try 

xm create <other options> vmid=X

That should work. So far I have only been testing with the -n option.
I haven't booted any domains yet. Hope this helps.

Niraj

On Fri, 9 Jul 2004 14:54:01 +0200, Rune Johan Andresen
<runejoha@idi.ntnu.no> wrote:
> 
> Is this a common problem for the unstable ver. or is this an exception
> for me and Niraj? ref: "xm create" - >
> (I am able to start xend without error msgs now) Maybe I should use the
> 1.2 ver for benchmarking, or does the
> unstable perform better?
> 
> Rune
> 
> 
[snip again]


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Mike Wray]

Keir Fraser wrote:

>>Clearer error messages than "Error: Cannot start domain" would probably be a
>>good start.  Anyone could have figured out in short order that the domain
>>couldn't be started.  What's much more important is *why* the domain
>>couldn't be started.

You should find that the error reason is being returned along with
the 'cannot start domain message now'. Up to now I've been letting
xend exceptions go uncaught so I can debug the resulting stacktrace,
so it's onlt recently that the errors have been caught and sent back
to the client (xm).

> This is absolutely true.
> 
> 
>>In general, error messages should always say why an operation failed, not
>>just that the operation failed.  I find that this works wonders in my own
>>programs.  In C programs, for example, always print out 'errno' in any error
>>message where it could possibly be relevant.  (Of course, to avoid
>>confusion, definitely don't print out errno where it's not valid anymore :))
>>The errno equivalent in perl is $!.  In python I guess you have to use
>>exceptions or something.

Errno is not usually that useful - it's usually something higher-level
that went wrong in most programs.

> 
> The problem with the new toolset is that some of the context for
> building a useful error message at the client program is separated
> from the client in both space (its in a server, behind a socket) and
> in time (the failure may be due to some random asynchronous request
> made ages ago). A tedious problem but certainly surmountable -- this
> area definitely needs attention.
> 
> The crux of the problem is that we need a domain controller capable of
> juggling multiple outstanding operations at the same time. In general
> you do that by using multiple threads or by decoupling requests from
> their eventual responses. Both can have their drawbacks, and
> asynchronous models in particular need great care to avoid runaway
> complexity.
> 
> My preferred model would be lightweight language-level threads, but I
> don't know if anything suitable exists for Python. What I've seen of
> Twisted so far hasn't recommended it to me. :-)

The problem is not with Twisted - it's with the way that domains
have to be built and shutdown. Twisted is merely a handy framework for
asynchronous programming - basically a big select loop.

Domain create used to be pretty simple - call create in xc with the
domain image, number of nics, disk spec and away you go.
It's now much more complex, involving at least 5 independent parties
connected by message streams: xen, xend, device driver domain(s), actual domain,
client (xm).

One a domain memory image is created, all devices have to be set up
using messaging from xend to the device driver domain(s) and the new domain.
All this is asynchronous, and not order-preserving, so it's intrinsically
fairly complex. And the messaging is all multiplexed over a single channel.
And it's mixed with the console i/o too.

Xend has to keep lots of state - like all the devices - otherwise the domain
can't even be shutdown. And since shutdown is intrinsically asynchronous this
is complicated too.

This could be programmed 2 ways: event-driven (asynch), or using threads.
I've done both in my time and each has its advantages and disadvantages.
Neither is easy.

In my view we need to do some work on the state kept by device drivers,
domains, and xen so that things can be simplified and xend doesn't
need to keep so much fragile state and order information.

Mike


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Keir Fraser]

> Keir Fraser wrote:
> 
> >>Clearer error messages than "Error: Cannot start domain" would probably be a
> >>good start.  Anyone could have figured out in short order that the domain
> >>couldn't be started.  What's much more important is *why* the domain
> >>couldn't be started.
> 
> You should find that the error reason is being returned along with
> the 'cannot start domain message now'. Up to now I've been letting
> xend exceptions go uncaught so I can debug the resulting stacktrace,
> so it's onlt recently that the errors have been caught and sent back
> to the client (xm).

Also the problems we've been seeing really have amounted to "internal
server errors" or similar unhelpful things. :-) Since the failures
have been due to bugs in xend and in the device drivers, it's not
possible to give a more useful error message to the user because the
problems weren't their fault.

> In my view we need to do some work on the state kept by device drivers,
> domains, and xen so that things can be simplified and xend doesn't
> need to keep so much fragile state and order information.

Maybe this can be cleaned up somewhat, but taking the example of i/o:
 Xen doesn't know about I/O any more (except carving up the PCI space).
 Device drivers are supposed to contain only 'soft state' so that they
 can easily be restarted.
 Maybe domains could do more, but they can't be trusted.

 -- Keir


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Ian Pratt]

> 
> Is this a common problem for the unstable ver. or is this an exception 
> for me and Niraj? ref: "xm create" - >
> (I am able to start xend without error msgs now) Maybe I should use the 
> 1.2 ver for benchmarking, or does the
> unstable perform better?

1.2 should be fine for benchmarking, and right now is certainly
the best choice if you want a hassle-free experience.

In the unstable tree, the 'xm' tool's command line interface has
skant documentation and is undergoing rapid revision. More so for
the various configuration files.

I hope we can release do a 2.0-alpha release in a week or so, in
which case we hope people will pitch in with testing and some
user documentation.

Thanks,
Ian


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Ian Pratt]

> Hi, after install all the necessary packages I also have problems to make
> everything function:
> 
>  brctl addbr nbe-br
> device nbe-br already exists; can't create bridge with the same name
> Interface not found: eth0

Rune,

Please can you try the latest repo.

If you look in the /etc/xen/xend-config.sxp configuration file
you'll see that there are two shell scripts which control the
networking configuration:

 * /etc/xen/network :

Run when xend starts, it creates bridges, adds physical
interfaces to bridges, and transfers IP addresses and routes from
physical interfaces to the bridge.  This script will work fine
for most people, with simple setups (e.g. one bridge, eth0 is the
sole physical interface), but if you want to do anything more
exciting (e.g. multiple phys interfaces or bridge, 802.3ab
channel bonding etc) then you'll need to edit this file. It also
takes a parameter 'vif-antispoof' which uses iptables to disable
forwarding by default, then adds specific rules for given src ip
addrs when domains are created (via /etc/xen/vif-bridge)

 * /etc/xen/vif-bridge

This is the default script that's run whenever a domain is
created to configure the virtual networking for the domain. It
adds the vif to the specified bridge, and can also be used to add
iptables rules to control forwarding of the domain's packets
through the bridge. 

We decided to switch to shell scripts to make it easier for
people to tailor them to their own config, rather than having to
hack python.

BTW: xend seems to be working quite stably for me these days. How
are other people finding it?

Thanks,
Ian



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

RE: still can't start domain

[James Harper]

[-- Attachment #1: Type: text/plain, Size: 1095 bytes --]

> BTW: xend seems to be working quite stably for me these days. How
> are other people finding it?

I checked out and built the latest copy earlier today and now dom0 won't boot. That's the first time that has happened in a few weeks, and probably only because I checked it out at an unlucky time!

Other than that, I have seen a few occasions where the whole server (xen + dom0...domN) have all rebooted, and another today where I logged in and found that dom1 had stopped. Dom1 is test of a backup mailserver, and due to the way spammers don't always use the primary MX record it gets a few messages, and does spam filtering. The beauty of using this as a test is that anything that goes via a secondary mailserver when the primary is up, is almost certainly spam and doesn't matter if it gets lost.

The server is in at the office and i'm at home high on painkillers after having my wisdom teeth extracted so checking it out will have to wait until tomorrow.

btw, what's the state of migration? I've seen some mentions of it in the changelogs but nothing on how to do it.

James

[-- Attachment #2: Type: text/html, Size: 1803 bytes --]

Re: still can't start domain

[Ian Pratt]

> > BTW: xend seems to be working quite stably for me these days. How
> > are other people finding it?
> 
> I checked out and built the latest copy earlier today and now dom0 won't boot. That's the first time that has happened in a few weeks, and probably only because I checked it out at an unlucky time!

Very strange.

I've just done a pull of xeno-unstable and 'make world', and it
boots fine for me, on two different machines.

The last significant checkin was 21h ago, so I don't think there
was a bad window. Please can you retry. Perhaps you tried booting
mis-matched xen and xenlinux?

> Other than that, I have seen a few occasions where the whole
> server (xen + dom0...domN) have all rebooted, and another today
> where I logged in and found that dom1 had stopped.

I've not seen a spontaneous reboot or domain crash in ages, ever
since the netif backend bug that was fixed.

Please can you try and get a serial line on the box to see if any
messages come out. Any sign of an oops message on the domain's
console?

> btw, what's the state of migration? I've seen some mentions of
> it in the changelogs but nothing on how to do it.

Live migration got broken by NGIO and the new tools, but it only
needs a couple of minor tweaks and it'll probably be working by
the end of the week.

Ian



-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Jody Belka]

*** Resending this on to the list, just for the record, now that
*** sourceforge email is working again

On Mon, Jul 12, 2004 at 09:27:36AM +0100, Keir Fraser wrote:
> Actually, I think I've found and fixed the bug. Please pull and test.
> Presumably you were using gcc-3.3.x? :-)

well, i'm actually using 2.95.x for the domain kernels, and 3.0.x for
xen (and only there because it won't build under 2.95.x). 

all working again with your bugfix though anyway, ta.


(nb, i'm guessing this won't actually make the list, seeing as sourceforge
is playing silly buggers with its email right now, but i'll try anyway)

-- 
Jody Belka
knew (at) pimb (dot) org


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: still can't start domain

[Jody Belka]

*** Resending this on to the list, just for the record, now that
*** sourceforge email is working again

On Sun, Jul 11, 2004 at 08:30:00PM +0100, Ian Pratt wrote:
> > I checked out and built the latest copy earlier today and now dom0 won't boot. That's the first time that has happened in a few weeks, and probably only because I checked it out at an unlucky time!
> 
> Very strange.
> 
> I've just done a pull of xeno-unstable and 'make world', and it
> boots fine for me, on two different machines.
> 
> The last significant checkin was 21h ago, so I don't think there
> was a bad window. Please can you retry. Perhaps you tried booting
> mis-matched xen and xenlinux?

I've found this as well on my system today.

Old rev = 40ed42de9JFsw0T591T5jpnZ82grCw (20040708124934)
New rev = 40f14d93VNnMM_iUKi37Qk5Dj84vwA (20040711142419)

If i try and boot using the xen from the new rev, i get the
following error on the screen. it doesn't print to the
serial console:

(XEN) Unknown interrupt


Booting a dom0 from either rev works with the xen from the old
rev, as does starting a domN once the system is booted.


-- 
Jody Belka
knew (at) pimb (dot) org


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

RE: still can't start domain

[James Harper]

The following was sent when the list was down. It's all working now with
the relevent update. Fyi, my gcc version is 'gcc (GCC) 3.3.4 (Debian)'
as reported by gcc --version.

Here's my original, no longer relevent, post:

My bad.

Someone had pulled out one of the power connectors to the redundant
power supplies, so on reboot it was waiting for someone to acknowledge
the fact that that the power subsystem was not redundant.

After hitting F1 to acknowledge though, xen said unknown interrupt and
refused to do anything else. This was before even loading dom0. This
still happened after I plugged up all the power supplies again and
rebooted.

James




-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
digital self defense, top technical experts, no vendor pitches,
unmatched networking opportunities. Visit www.blackhat.com