can't start domains

can't start domains

[James Harper]

[-- Attachment #1: Type: text/plain, Size: 2724 bytes --]

i've just set up xen on a new machine and now get an error when trying to start a domain. It goes like this:

op_create> Exception creating domain:
Traceback (most recent call last):
  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDomainDir.py", line 64, in op_create
    deferred = self.xd.domain_create(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomain.py", line 359, in domain_create
    deferred = XendDomainInfo.vm_create(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 259, in vm_create
    return vm.construct(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 511, in construct
    self.destroy()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 613, in destroy
    self.cleanup()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 641, in cleanup
    self.release_devices()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 652, in release_devices
    self.release_vbds()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 675, in release_vbds
    ctrl.destroy()
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 364, in destroy
    self.destroyDevices()
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 371, in destroyDevices
    dev.destroy()
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 207, in destroy
    log.debug("Destroying vbd domain=%d vdev=%d", self.controller.dom, self.vdev)
  File "/usr/lib/python2.3/logging/__init__.py", line 879, in debug
    apply(self._log, (DEBUG, msg, args), kwargs)
  File "/usr/lib/python2.3/logging/__init__.py", line 994, in _log
    self.handle(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 1004, in handle
    self.callHandlers(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 1037, in callHandlers
    hdlr.handle(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 592, in handle
    self.emit(record)
  File "/usr/lib/python2.3/logging/handlers.py", line 102, in emit
    msg = "%s\n" % self.format(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 567, in format
    return fmt.format(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 362, in format
    record.message = record.getMessage()
  File "/usr/lib/python2.3/logging/__init__.py", line 233, in getMessage
    msg = msg % self.args
TypeError: int argument required
Error: Error creating domain: int argument required

it got to be something i haven't installed... maybe it's too late at night... any ideas?

thanks

James

[-- Attachment #2: Type: text/html, Size: 3934 bytes --]

Re: can't start domains

[Andrew Warfield]

I suspect this crash is just a glitch in the python, but that your
domain isn't starting for some other reason.  Try commenting out the
line:

  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py",
line 207, in destroy
    log.debug("Destroying vbd domain=%d vdev=%d", self.controller.dom,
self.vdev)

and see if it gets any further with the crash.

a.


----- Original Message -----
From: James Harper <jamesh@bendigoit.com.au>
Date: Sat, 21 Aug 2004 22:53:19 +1000
Subject: [Xen-devel] can't start domains
To: "xen-devel@lists.sourceforge.net" <xen-devel@lists.sourceforge.net>

 
i've just set up xen on a new machine and now get an error when trying
to start a domain. It goes like this:
  
op_create> Exception creating domain:
Traceback (most recent call last):
  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDomainDir.py",
line 64, in op_create
    deferred = self.xd.domain_create(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomain.py", line
359, in domain_create
    deferred = XendDomainInfo.vm_create(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py",
line 259, in vm_create
    return vm.construct(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py",
line 511, in construct
    self.destroy()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py",
line 613, in destroy
    self.cleanup()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py",
line 641, in cleanup
    self.release_devices()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py",
line 652, in release_devices
    self.release_vbds()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py",
line 675, in release_vbds
    ctrl.destroy()
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py",
line 364, in destroy
    self.destroyDevices()
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py",
line 371, in destroyDevices
    dev.destroy()
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py",
line 207, in destroy
    log.debug("Destroying vbd domain=%d vdev=%d", self.controller.dom,
self.vdev)
  File "/usr/lib/python2.3/logging/__init__.py", line 879, in debug
    apply(self._log, (DEBUG, msg, args), kwargs)
  File "/usr/lib/python2.3/logging/__init__.py", line 994, in _log
    self.handle(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 1004, in handle
    self.callHandlers(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 1037, in callHandlers
    hdlr.handle(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 592, in handle
    self.emit(record)
  File "/usr/lib/python2.3/logging/handlers.py", line 102, in emit
    msg = "%s\n" % self.format(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 567, in format
    return fmt.format(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 362, in format
    record.message = record.getMessage()
  File "/usr/lib/python2..3/logging/__init__.py", line 233, in getMessage
    msg = msg % self.args
TypeError: int argument required
Error: Error creating domain: int argument required
 
it got to be something i haven't installed.... maybe it's too late at
night... any ideas?
  
thanks 
  
James


-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

Re: can't start domains

[Ian Pratt]

Something is going wrong during the domain creation, and then
xend tries to clean up by destroying the stillborn
domain. Unfortunately, we then hit a xend bug while destroying
the domain's block device (perhaps because that was the original
problem?) and the traceback all refers to this bug which prevents
it from reporting back to xm what the original problem was.

I've cc'ed Mike so that he can take a look at the Xend bug. 

As regards the real error, take a close look at the sxp config
data that that xm sends to xend and check that it matches your
expected configuration. (see the config: output from xend).

Ian


> i've just set up xen on a new machine and now get an error when trying to start a domain. It goes like this:
> 
> op_create> Exception creating domain:
> Traceback (most recent call last):
>   File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDomainDir.py", line 64, in op_create
>     deferred = self.xd.domain_create(config)
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomain.py", line 359, in domain_create
>     deferred = XendDomainInfo.vm_create(config)
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 259, in vm_create
>     return vm.construct(config)
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 511, in construct
>     self.destroy()
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 613, in destroy
>     self.cleanup()
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 641, in cleanup
>     self.release_devices()
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 652, in release_devices
>     self.release_vbds()
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 675, in release_vbds
>     ctrl.destroy()
>   File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 364, in destroy
>     self.destroyDevices()
>   File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 371, in destroyDevices
>     dev.destroy()
>   File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 207, in destroy
>     log.debug("Destroying vbd domain=%d vdev=%d", self.controller.dom, self.vdev)
>   File "/usr/lib/python2.3/logging/__init__.py", line 879, in debug
>     apply(self._log, (DEBUG, msg, args), kwargs)
>   File "/usr/lib/python2.3/logging/__init__.py", line 994, in _log
>     self.handle(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 1004, in handle
>     self.callHandlers(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 1037, in callHandlers
>     hdlr.handle(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 592, in handle
>     self.emit(record)
>   File "/usr/lib/python2.3/logging/handlers.py", line 102, in emit
>     msg = "%s\n" % self.format(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 567, in format
>     return fmt.format(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 362, in format
>     record.message = record.getMessage()
>   File "/usr/lib/python2.3/logging/__init__.py", line 233, in getMessage
>     msg = msg % self.args
> TypeError: int argument required
> Error: Error creating domain: int argument required
> 
> it got to be something i haven't installed... maybe it's too late at night... any ideas?



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

RE: can't start domains

[James Harper]

[-- Attachment #1: Type: text/plain, Size: 10052 bytes --]

the output i sent yesterday was a combination of xend and xm, here's just the xm output:

xen2:/etc/xen# xm create -f /etc/xen/gaia2
Using config file "/etc/xen/gaia2".
Error: Error creating domain: int argument required

here's /etc/xen/gaia2

#  -*- mode: python; -*-
#import sys
#import xenctl.ip
vmid=3
kernel = "/boot/vmlinuz-2.6.7-xenU"
memory = 128
name = "gaia2"
cpu = vmid
nics = 2
vif = [ 'mac=FE:FD:01:00:04:00,bridge=br0', 'mac=FE:FD:01:00:04:01,bridge=br1' ]
disk = [ 'phy:vg00/gaia2,sda1,w', 'phy:vg00/gaia2-swap,sda2,w' ]
root = "/dev/sda1 ro"
extra = " 4 VMID=%d " % vmid
restart = False

it is identical to the one that works on the other server, except that on this one I had to comment out the two include lines, as it couldn't find xenctl.ip, i don't think that is required anymore anyway is it? it's not in the example configs.

there must be something different between this new server and the other one... i just can't seem to figure out what though!!!

here's a more complete output from xend, i think i missed a bit in my last post:

config: (vm (name gaia2) (memory 128) (cpu 3) (image (linux (kernel /boot/vmlinuz-2.6.7-xenU) (root '/dev/sda1 ro') (args ' 4 VMID=3 '))) (device (vbd (uname phy:vg00/gaia2) (dev sda1) (mode w))) (device (vbd (uname phy:vg00/gaia2-swap) (dev sda2) (mode w))) (device (vif (mac FE:FD:01:00:04:00) (bridge br0))) (device (vif (mac FE:FD:01:00:04:01) (bridge br1))))
[xend] DEBUG (XendDomainInfo:717) init_domain> Created domain=1 name=gaia2 memory=128
[xend] INFO (console:93) Created console id=14 domain=1 port=9601
VIRTUAL MEMORY ARRANGEMENT:
 Loaded kernel: c0100000->c02fbf44
 Init. ramdisk: c02fc000->c02fc000
 Phys-Mach map: c02fc000->c031c000
 Page tables:   c031c000->c031e000
 Start info:    c031e000->c031f000
 Boot stack:    c031f000->c0320000
 TOTAL:         c0100000->c0400000
 ENTRY ADDRESS: c0100000
[xend] DEBUG (XendDomainInfo:1065) Creating vbd dom=1 uname=phy:vg00/gaia2 dev=sda1
[xend] DEBUG (blkif:50) Connecting blkif <BlkifBackendController 1 0>
CtrlMsgRcvr>writeRequest> message: type= blkif_be_create_t 1:0 id=1
payload= {'status': 1852731203, 'blkif_handle': 0, 'domid': 1}
[xend] DEBUG (XendDomainInfo:1065) Creating vbd dom=1 uname=phy:vg00/gaia2-swap dev=sda2
[xend] DEBUG (XendDomainInfo:674) Destroying vbds for domain 1
[xend] DEBUG (blkif:363) Destroying blkif domain=1
op_create> Exception creating domain:
Traceback (most recent call last):
  File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDomainDir.py", line 64, in op_create
    deferred = self.xd.domain_create(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomain.py", line 359, in domain_create
    deferred = XendDomainInfo.vm_create(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 259, in vm_create
    return vm.construct(config)
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 511, in construct
    self.destroy()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 613, in destroy
    self.cleanup()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 641, in cleanup
    self.release_devices()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 652, in release_devices
    self.release_vbds()
  File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 675, in release_vbds
    ctrl.destroy()
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 364, in  destroy
    self.destroyDevices()
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 371, in destroyDevices
    dev.destroy()
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 207, in destroy
    log.debug("Destroying vbd domain=%d vdev=%d", self.controller.dom, self.vdev
)
  File "/usr/lib/python2.3/logging/__init__.py", line 879, in debug
    apply(self._log, (DEBUG, msg, args), kwargs)
  File "/usr/lib/python2.3/logging/__init__.py", line 994, in _log
    self.handle(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 1004, in handle
    self.callHandlers(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 1037, in callHandlers
    hdlr.handle(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 592, in handle
    self.emit(record)
  File "/usr/lib/python2.3/logging/handlers.py", line 102, in emit
    msg = "%s\n" % self.format(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 567, in format
    return fmt.format(record)
  File "/usr/lib/python2.3/logging/__init__.py", line 362, in format
    record.message = record.getMessage()
  File "/usr/lib/python2.3/logging/__init__.py", line 233, in getMessage
    msg = msg % self.args
TypeError: int argument required
EVENT> xend.console.create [14, 1, 9601]
requestReceived> message: type= None 1:32 id=255
payload= {'status': 1}
requestReceived> No device: Message type netif_be_driver_status_changed_t 3:32 < Channel dom=0 ports=2:-1>
responseReceived> message: type= blkif_be_create_t 1:0 id=1 payload= {'status': 0, 'blkif_handle': 0, 'domid': 1}
respond_be_create> {'status': 0, 'blkif_handle': 0, 'domid': 1}
Unhandled error in Deferred:
Traceback (most recent call last):
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 313, in _startRunCallbacks
    self._runCallbacks()
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
    self.result = callback(self.result, *args, **kw)
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 252, in callback
    self._startRunCallbacks(result)
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 313, in _startRunCallbacks
    self._runCallbacks()
--- <exception caught here> ---
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
    self.result = callback(self.result, *args, **kw)
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 226, in send_be_vbd_create
    { 'domid'        : self.controller.dom,
  File "/usr/lib/python2.3/site-packages/xen/xend/server/messages.py", line 217, in packMsg
    msg = xu.message(major, minor, msgid, args)
exceptions.TypeError: payload contains bad items

thanks

James



From: Ian Pratt
Sent: Sat 21/08/2004 11:21 PM
To: James Harper
Cc: xen-devel@lists.sourceforge.net; Ian.Pratt@cl.cam.ac.uk; mike.wray@hpl.hp.com
Subject: Re: [Xen-devel] can't start domains


Something is going wrong during the domain creation, and then
xend tries to clean up by destroying the stillborn
domain. Unfortunately, we then hit a xend bug while destroying
the domain's block device (perhaps because that was the original
problem?) and the traceback all refers to this bug which prevents
it from reporting back to xm what the original problem was.

I've cc'ed Mike so that he can take a look at the Xend bug. 

As regards the real error, take a close look at the sxp config
data that that xm sends to xend and check that it matches your
expected configuration. (see the config: output from xend).

Ian


> i've just set up xen on a new machine and now get an error when trying to start a domain. It goes like this:
> 
> op_create> Exception creating domain:
> Traceback (most recent call last):
>   File "/usr/lib/python2.3/site-packages/xen/xend/server/SrvDomainDir.py", line 64, in op_create
>     deferred = self.xd.domain_create(config)
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomain.py", line 359, in domain_create
>     deferred = XendDomainInfo.vm_create(config)
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 259, in vm_create
>     return vm.construct(config)
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 511, in construct
>     self.destroy()
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 613, in destroy
>     self.cleanup()
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 641, in cleanup
>     self.release_devices()
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 652, in release_devices
>     self.release_vbds()
>   File "/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py", line 675, in release_vbds
>     ctrl.destroy()
>   File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 364, in destroy
>     self.destroyDevices()
>   File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 371, in destroyDevices
>     dev.destroy()
>   File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 207, in destroy
>     log.debug("Destroying vbd domain=%d vdev=%d", self.controller.dom, self.vdev)
>   File "/usr/lib/python2.3/logging/__init__.py", line 879, in debug
>     apply(self._log, (DEBUG, msg, args), kwargs)
>   File "/usr/lib/python2.3/logging/__init__.py", line 994, in _log
>     self.handle(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 1004, in handle
>     self.callHandlers(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 1037, in callHandlers
>     hdlr.handle(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 592, in handle
>     self.emit(record)
>   File "/usr/lib/python2.3/logging/handlers.py", line 102, in emit
>     msg = "%s\n" % self.format(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 567, in format
>     return fmt.format(record)
>   File "/usr/lib/python2.3/logging/__init__.py", line 362, in format
>     record.message = record.getMessage()
>   File "/usr/lib/python2.3/logging/__init__.py", line 233, in getMessage
>     msg = msg % self.args
> TypeError: int argument required
> Error: Error creating domain: int argument required
> 
> it got to be something i haven't installed... maybe it's too late at night... any ideas?

[-- Attachment #2: Type: text/html, Size: 12809 bytes --]

Re: can't start domains

[Ian Pratt]

> the output i sent yesterday was a combination of xend and xm, here's just the xm output:
> 
> xen2:/etc/xen# xm create -f /etc/xen/gaia2
> Using config file "/etc/xen/gaia2".
> Error: Error creating domain: int argument required
> 
> here's /etc/xen/gaia2
> 
> #  -*- mode: python; -*-
> #import sys
> #import xenctl.ip
> vmid=3
> kernel = "/boot/vmlinuz-2.6.7-xenU"
> memory = 128
> name = "gaia2"
> cpu = vmid
> nics = 2
> vif = [ 'mac=FE:FD:01:00:04:00,bridge=br0', 'mac=FE:FD:01:00:04:01,bridge=br1' ]
> disk = [ 'phy:vg00/gaia2,sda1,w', 'phy:vg00/gaia2-swap,sda2,w' ]
> root = "/dev/sda1 ro"
> extra = " 4 VMID=%d " % vmid
> restart = False


Not withstanding the bug in Xend's error reporting path, it looks
like something is going wrong exporting vg00/gaia2-swap.

Can you do an 'sfdisk -s /dev/vg00/gaia2-swap' to check that dom0 is
happy with it.

The next thing to do would be to enable the debug printf's in
drivers/xen/blkback/common.h and then you might get some useful
messages on the dom0 console ('dmesg').

BTW: There was a multiple NICs bug that got fixed recently, so
you should probably be using a very recent checkout.


Ian


-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

RE: can't start domains

[James Harper]

[-- Attachment #1: Type: text/plain, Size: 2779 bytes --]

i just realised that the other difference between the working server and the not working server is that the working one is using 2.4 in dom0 (and the 2.4 version of lvm) while the non-working server is using 2.6 in dom0 (and thus the 2.6 version of lvm).

dom0 is happy to mkswap on /dev/vg00/gaia2-swap, and to mount /dev/vg00/gaia2.

those are symlinks to /dev/mapper, and I have tried mounting the files that the symlink points to without any success, in case that matters.

If I remove the "disk =" line out of the config file, then it creates the domain, which in turn gets as far as wanting to mount the root filesystem, a lot more success.

If I change the disk line to use my swap partition instead (the only other partition I can usefully point to), it still breaks in the same way as when I was using /dev/vg00/gaia2

So for some reason, I have a problem using any block device from xen!!!

Removing the nics from my config didn't make any difference. fwiw, i've always used multiple nics and have never had a problem with them. Maybe i just missed the bad spot.

One mistake I made earlier was having some old 1.2 files lying around from when I tested this server a while ago. I think I have cleaned them up, but apart from the possibility of 2.6 causing a problem, that's my most likely cause. Where should I look for stale files?

thanks

James



From: Ian Pratt
Sent: Sun 22/08/2004 6:24 PM
To: James Harper
Cc: Ian Pratt; xen-devel@lists.sourceforge.net; mike.wray@hpl.hp.com; Ian.Pratt@cl.cam.ac.uk
Subject: Re: [Xen-devel] can't start domains


> the output i sent yesterday was a combination of xend and xm, here's just the xm output:
> 
> xen2:/etc/xen# xm create -f /etc/xen/gaia2
> Using config file "/etc/xen/gaia2".
> Error: Error creating domain: int argument required
> 
> here's /etc/xen/gaia2
> 
> #  -*- mode: python; -*-
> #import sys
> #import xenctl.ip
> vmid=3
> kernel = "/boot/vmlinuz-2.6.7-xenU"
> memory = 128
> name = "gaia2"
> cpu = vmid
> nics = 2
> vif = [ 'mac=FE:FD:01:00:04:00,bridge=br0', 'mac=FE:FD:01:00:04:01,bridge=br1' ]
> disk = [ 'phy:vg00/gaia2,sda1,w', 'phy:vg00/gaia2-swap,sda2,w' ]
> root = "/dev/sda1 ro"
> extra = " 4 VMID=%d " % vmid
> restart = False


Not withstanding the bug in Xend's error reporting path, it looks
like something is going wrong exporting vg00/gaia2-swap.

Can you do an 'sfdisk -s /dev/vg00/gaia2-swap' to check that dom0 is
happy with it.

The next thing to do would be to enable the debug printf's in
drivers/xen/blkback/common.h and then you might get some useful
messages on the dom0 console ('dmesg').

BTW: There was a multiple NICs bug that got fixed recently, so
you should probably be using a very recent checkout.


Ian

[-- Attachment #2: Type: text/html, Size: 4120 bytes --]

Re: can't start domains

[Ian Pratt]

 
> If I change the disk line to use my swap partition instead (the only other partition I can usefully point to), it still breaks in the same way as when I was using /dev/vg00/gaia2

If I understand correctly, you're saying it breaks in the same
way (with a similar python traceback) even if you have the
simplest possible setup of a single VBD exporting a physical
partition.

This means we can ignore the LVM issue.

Can you post the traceback for this simple case, and also the
debug output from dom0 having enabled the debug prints in blkback/common.h

> So for some reason, I have a problem using any block device from xen!!!

I think we'd have noticed this if it were a general problem ;-)

Are you fully up to date with your repo? There have been a few
duff ones in the last couple of weeks, but things seem stable
now.  We came very close to putting a "Xen 2.0-rc1" designation on
the tree Friday afternoon, but thought we leave a few tests
running over the weekend.

I just noticed that 2.4.27 and 2.6.8 have just come out, so I
guess we should upgrade before declaring the 2.0 release
candidate. Oh well...

> One mistake I made earlier was having some old 1.2 files lying around from when I tested this server a while ago. I think I have cleaned them up, but apart from the possibility of 2.6 causing a problem, that's my most likely cause. Where should I look for stale files?

'make uninstall' is fairly brutal but should clear things out.

Ian


-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

RE: can't start domains

[James Harper]

[-- Attachment #1: Type: text/plain, Size: 6218 bytes --]

make uninstall does this:
        cp -a /etc/xen /etc/xen.old && rm -rf /etc/xen
        rm -rf "/usr/lib/python2.2/site-packages/xen* /usr/lib/libxc* /usr/lib/python2.2/site-packages/Xc*"

which might make more sense if i wasn't using python 2.3. also, there's something funny going on with /usr/lib/libxc, I get:
libxc.so
libxc.so.1.3
libxc.so.1.3.0

I think the first two should be symlinks to the third?

but that's irrelevant to my situation. you'll see in the output below, the error occurs in /usr/lib/python2.3/site-packages/xen/xend/server/blkif.py:228. At this point, backend.handle = 0. is that right?

i'll recompile with the debug flag set shortly and send the output in the morning (AEST)

here's the output:

config: (vm (name gaia2) (memory 128) (cpu 3) (image (linux (kernel /boot/vmlinuz-2.6.7-xenU) (root '/dev/sda1 ro') (args ' 4 VMID=3 '))) (device (vbd (uname phy:ida/c0d0p5) (dev sda1) (mode w))))
[xend] DEBUG (XendDomainInfo:726) init_domain> Created domain=4 name=gaia2 memory=128
[xend] INFO (console:93) Created console id=17 domain=4 port=9604
VIRTUAL MEMORY ARRANGEMENT:
 Loaded kernel: c0100000->c02fbf44
 Init. ramdisk: c02fc000->c02fc000
 Phys-Mach map: c02fc000->c031c000
 Page tables:   c031c000->c031e000
 Start info:    c031e000->c031f000
 Boot stack:    c031f000->c0320000
 TOTAL:         c0100000->c0400000
 ENTRY ADDRESS: c0100000
[xend] DEBUG (blkif:52) Connecting blkif <BlkifBackendController 4 0>
CtrlMsgRcvr>writeRequest> message: type= blkif_be_create_t 1:0 id=1
payload= {'status': 1852731203, 'blkif_handle': 0, 'domid': 4}
EVENT> xend.console.create [17, 4, 9604]
responseReceived> message: type= blkif_be_create_t 1:0 id=1
payload= {'status': 0, 'blkif_handle': 0, 'domid': 4}
respond_be_create> {'status': 0, 'blkif_handle': 0, 'domid': 4}
[xend] DEBUG (XendDomainInfo:1083) Creating vbd dom=4 uname=phy:ida/c0d0p5 dev=sda1
[xend] DEBUG (blkif:52) Connecting blkif <BlkifBackendController 4 0>
Unhandled error in Deferred:
Traceback (most recent call last):
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 205, in addCallbacks
    self._runCallbacks()
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
    self.result = callback(self.result, *args, **kw)
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 252, in callback
    self._startRunCallbacks(result)
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 313, in _startRunCallbacks
    self._runCallbacks()
--- <exception caught here> ---
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
    self.result = callback(self.result, *args, **kw)
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 228, in send_be_vbd_create
    { 'domid'        : self.controller.dom,
  File "/usr/lib/python2.3/site-packages/xen/xend/server/messages.py", line 217, in packMsg
    msg = xu.message(major, minor, msgid, args)
exceptions.TypeError: payload contains bad items
[xend] DEBUG (XendDomainInfo:680) Destroying vbds for domain 4
[xend] DEBUG (blkif:364) Destroying blkif domain=4
op_create> Deferred Exception creating domain: [Failure instance: Traceback: exceptions.TypeError, int argument required
/usr/lib/python2.3/site-packages/twisted/internet/defer.py:338:_runCallbacks
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:512:cberr
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:619:destroy
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:647:cleanup
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:658:release_devices
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:681:release_vbds
/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py:365:destroy
/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py:372:destroyDevices
/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py:209:destroy
/usr/lib/python2.3/logging/__init__.py:879:debug
/usr/lib/python2.3/logging/__init__.py:994:_log
/usr/lib/python2.3/logging/__init__.py:1004:handle
/usr/lib/python2.3/logging/__init__.py:1037:callHandlers
/usr/lib/python2.3/logging/__init__.py:592:handle
/usr/lib/python2.3/logging/handlers.py:102:emit
/usr/lib/python2.3/logging/__init__.py:567:format
/usr/lib/python2.3/logging/__init__.py:362:format
/usr/lib/python2.3/logging/__init__.py:233:getMessage
]


















From: Ian Pratt
Sent: Sun 22/08/2004 8:33 PM
To: James Harper
Cc: Ian Pratt; xen-devel@lists.sourceforge.net; mike.wray@hpl.hp.com; Ian.Pratt@cl.cam.ac.uk
Subject: Re: [Xen-devel] can't start domains


 
> If I change the disk line to use my swap partition instead (the only other partition I can usefully point to), it still breaks in the same way as when I was using /dev/vg00/gaia2

If I understand correctly, you're saying it breaks in the same
way (with a similar python traceback) even if you have the
simplest possible setup of a single VBD exporting a physical
partition.

This means we can ignore the LVM issue.

Can you post the traceback for this simple case, and also the
debug output from dom0 having enabled the debug prints in blkback/common.h

> So for some reason, I have a problem using any block device from xen!!!

I think we'd have noticed this if it were a general problem ;-)

Are you fully up to date with your repo? There have been a few
duff ones in the last couple of weeks, but things seem stable
now.  We came very close to putting a "Xen 2.0-rc1" designation on
the tree Friday afternoon, but thought we leave a few tests
running over the weekend.

I just noticed that 2.4.27 and 2.6.8 have just come out, so I
guess we should upgrade before declaring the 2.0 release
candidate. Oh well...

> One mistake I made earlier was having some old 1.2 files lying around from when I tested this server a while ago. I think I have cleaned them up, but apart from the possibility of 2.6 causing a problem, that's my most likely cause. Where should I look for stale files?

'make uninstall' is fairly brutal but should clear things out.

Ian

[-- Attachment #2: Type: text/html, Size: 8580 bytes --]

RE: can't start domains

[James Harper]

[-- Attachment #1: Type: text/plain, Size: 6950 bytes --]

I figured it out, at line 228 self.vdev = None (not an integer!!!), because I don't have a /dev/sda1 in /dev so xen can't figure out the major & minor numbers. I run udev which removes anything not being used. My scsi controller is an (old & slow) compaq smart array and the devices for that live in /dev/ida.

I believe this is a bug, but how would you solve it? The fact that the error isn't reported nicely is definitely a bug though, imho.

well... at least i have it working again now. Anyone care to add this to the FAQ?

James



From: James Harper
Sent: Sun 22/08/2004 9:18 PM
To: Ian Pratt
Cc: Ian Pratt; xen-devel@lists.sourceforge.net; mike.wray@hpl.hp.com
Subject: RE: [Xen-devel] can't start domains


make uninstall does this:
        cp -a /etc/xen /etc/xen.old && rm -rf /etc/xen
        rm -rf "/usr/lib/python2.2/site-packages/xen* /usr/lib/libxc* /usr/lib/python2.2/site-packages/Xc*"

which might make more sense if i wasn't using python 2.3. also, there's something funny going on with /usr/lib/libxc, I get:
libxc.so
libxc.so.1.3
libxc.so.1.3.0

I think the first two should be symlinks to the third?

but that's irrelevant to my situation. you'll see in the output below, the error occurs in /usr/lib/python2.3/site-packages/xen/xend/server/blkif.py:228. At this point, backend.handle = 0. is that right?

i'll recompile with the debug flag set shortly and send the output in the morning (AEST)

here's the output:

config: (vm (name gaia2) (memory 128) (cpu 3) (image (linux (kernel /boot/vmlinuz-2.6.7-xenU) (root '/dev/sda1 ro') (args ' 4 VMID=3 '))) (device (vbd (uname phy:ida/c0d0p5) (dev sda1) (mode w))))
[xend] DEBUG (XendDomainInfo:726) init_domain> Created domain=4 name=gaia2 memory=128
[xend] INFO (console:93) Created console id=17 domain=4 port=9604
VIRTUAL MEMORY ARRANGEMENT:
 Loaded kernel: c0100000->c02fbf44
 Init. ramdisk: c02fc000->c02fc000
 Phys-Mach map: c02fc000->c031c000
 Page tables:   c031c000->c031e000
 Start info:    c031e000->c031f000
 Boot stack:    c031f000->c0320000
 TOTAL:         c0100000->c0400000
 ENTRY ADDRESS: c0100000
[xend] DEBUG (blkif:52) Connecting blkif <BlkifBackendController 4 0>
CtrlMsgRcvr>writeRequest> message: type= blkif_be_create_t 1:0 id=1
payload= {'status': 1852731203, 'blkif_handle': 0, 'domid': 4}
EVENT> xend.console.create [17, 4, 9604]
responseReceived> message: type= blkif_be_create_t 1:0 id=1
payload= {'status': 0, 'blkif_handle': 0, 'domid': 4}
respond_be_create> {'status': 0, 'blkif_handle': 0, 'domid': 4}
[xend] DEBUG (XendDomainInfo:1083) Creating vbd dom=4 uname=phy:ida/c0d0p5 dev=sda1
[xend] DEBUG (blkif:52) Connecting blkif <BlkifBackendController 4 0>
Unhandled error in Deferred:
Traceback (most recent call last):
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 205, in addCallbacks
    self._runCallbacks()
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
    self.result = callback(self.result, *args, **kw)
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 252, in callback
    self._startRunCallbacks(result)
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 313, in _startRunCallbacks
    self._runCallbacks()
--- <exception caught here> ---
  File "/usr/lib/python2.3/site-packages/twisted/internet/defer.py", line 338, in _runCallbacks
    self.result = callback(self.result, *args, **kw)
  File "/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py", line 228, in send_be_vbd_create
    { 'domid'        : self.controller.dom,
  File "/usr/lib/python2.3/site-packages/xen/xend/server/messages.py", line 217, in packMsg
    msg = xu.message(major, minor, msgid, args)
exceptions.TypeError: payload contains bad items
[xend] DEBUG (XendDomainInfo:680) Destroying vbds for domain 4
[xend] DEBUG (blkif:364) Destroying blkif domain=4
op_create> Deferred Exception creating domain: [Failure instance: Traceback: exceptions.TypeError, int argument required
/usr/lib/python2.3/site-packages/twisted/internet/defer.py:338:_runCallbacks
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:512:cberr
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:619:destroy
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:647:cleanup
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:658:release_devices
/usr/lib/python2.3/site-packages/xen/xend/XendDomainInfo.py:681:release_vbds
/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py:365:destroy
/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py:372:destroyDevices
/usr/lib/python2.3/site-packages/xen/xend/server/blkif.py:209:destroy
/usr/lib/python2.3/logging/__init__.py:879:debug
/usr/lib/python2.3/logging/__init__.py:994:_log
/usr/lib/python2.3/logging/__init__.py:1004:handle
/usr/lib/python2.3/logging/__init__.py:1037:callHandlers
/usr/lib/python2.3/logging/__init__.py:592:handle
/usr/lib/python2.3/logging/handlers.py:102:emit
/usr/lib/python2.3/logging/__init__.py:567:format
/usr/lib/python2.3/logging/__init__.py:362:format
/usr/lib/python2.3/logging/__init__.py:233:getMessage
]


















From: Ian Pratt
Sent: Sun 22/08/2004 8:33 PM
To: James Harper
Cc: Ian Pratt; xen-devel@lists.sourceforge.net; mike.wray@hpl.hp.com; Ian.Pratt@cl.cam.ac.uk
Subject: Re: [Xen-devel] can't start domains


 
> If I change the disk line to use my swap partition instead (the only other partition I can usefully point to), it still breaks in the same way as when I was using /dev/vg00/gaia2

If I understand correctly, you're saying it breaks in the same
way (with a similar python traceback) even if you have the
simplest possible setup of a single VBD exporting a physical
partition.

This means we can ignore the LVM issue.

Can you post the traceback for this simple case, and also the
debug output from dom0 having enabled the debug prints in blkback/common.h

> So for some reason, I have a problem using any block device from xen!!!

I think we'd have noticed this if it were a general problem ;-)

Are you fully up to date with your repo? There have been a few
duff ones in the last couple of weeks, but things seem stable
now.  We came very close to putting a "Xen 2.0-rc1" designation on
the tree Friday afternoon, but thought we leave a few tests
running over the weekend.

I just noticed that 2.4.27 and 2.6.8 have just come out, so I
guess we should upgrade before declaring the 2.0 release
candidate. Oh well...

> One mistake I made earlier was having some old 1.2 files lying around from when I tested this server a while ago. I think I have cleaned them up, but apart from the possibility of 2.6 causing a problem, that's my most likely cause. Where should I look for stale files?

'make uninstall' is fairly brutal but should clear things out.

Ian

[-- Attachment #2: Type: text/html, Size: 9875 bytes --]

Re: can't start domains

[Ian Pratt]

> I figured it out, at line 228 self.vdev = None (not an integer!!!), because I don't have a /dev/sda1 in /dev so xen can't figure out the major & minor numbers. I run udev which removes anything not being used. My scsi controller is an (old & slow) compaq smart array and the devices for that live in /dev/ida.
> 
> I believe this is a bug, but how would you solve it? The fact that the error isn't reported nicely is definitely a bug though, imho.

Thanks for tracking this down. Xend certainly needs to handle the
error case more gracefully.

We're actually aware of a general problem with the way we handle
looking up block device names. Currently, xend looks both the
subject and target names up in domain 0, and then communicates
device numbers to the front and backend drivers.

This doesn't make sense if the guest domain is actually a NetBSD
OS with entirely different device names/numbers. It also doesn't
always work if you're using backend drivers in separate domains
from dom0 (to enable restartable device drivers), where you might
not have the subject device node in domain 0.

The correct fix is to look both the subject and target device
names up in the context of the domain that's going to be using
them. This will require a few changes to the device control
message protocol. It's on the todo list...

BTW: You can specify devices using the 16bit hex device
number. e.g. in your example to have a target of sda1:
  phy:vg/foobar,0803,w

Ian




-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

RE: can't start domains

[James Harper]

[-- Attachment #1: Type: text/plain, Size: 2045 bytes --]

the hex solution will work nicely. Maybe xen needs its own set of major/minor numbers which xend can parse without having to look them up? eg /dev/xbd1

I wonder how udev will work running in the xenU domain... do you think it would recognise the xen devices? i'll give it ago when I can afford some more testing time.

thanks

James



From: Ian Pratt
Sent: Mon 23/08/2004 2:37 AM
To: James Harper
Cc: Ian Pratt; xen-devel@lists.sourceforge.net; mike.wray@hpl.hp.com; Ian.Pratt@cl.cam.ac.uk
Subject: Re: [Xen-devel] can't start domains


> I figured it out, at line 228 self.vdev = None (not an integer!!!), because I don't have a /dev/sda1 in /dev so xen can't figure out the major & minor numbers. I run udev which removes anything not being used. My scsi controller is an (old & slow) compaq smart array and the devices for that live in /dev/ida.
> 
> I believe this is a bug, but how would you solve it? The fact that the error isn't reported nicely is definitely a bug though, imho.

Thanks for tracking this down. Xend certainly needs to handle the
error case more gracefully.

We're actually aware of a general problem with the way we handle
looking up block device names. Currently, xend looks both the
subject and target names up in domain 0, and then communicates
device numbers to the front and backend drivers.

This doesn't make sense if the guest domain is actually a NetBSD
OS with entirely different device names/numbers. It also doesn't
always work if you're using backend drivers in separate domains
from dom0 (to enable restartable device drivers), where you might
not have the subject device node in domain 0.

The correct fix is to look both the subject and target device
names up in the context of the domain that's going to be using
them. This will require a few changes to the device control
message protocol. It's on the todo list...

BTW: You can specify devices using the 16bit hex device
number. e.g. in your example to have a target of sda1:
  phy:vg/foobar,0803,w

Ian

[-- Attachment #2: Type: text/html, Size: 2705 bytes --]

Re: can't start domains

[Rik van Riel]

On Sun, 22 Aug 2004, Ian Pratt wrote:

> > So for some reason, I have a problem using any block device from xen!!!
> 
> I think we'd have noticed this if it were a general problem ;-)

I'm seeing the same problems with /dev/loop0 exported as
a VBD.

> Are you fully up to date with your repo? There have been a few
> duff ones in the last couple of weeks, but things seem stable
> now.

I'll upgrade the Xen utilities and xend tomorrow.  Hope
I didn't spend a few days trying to debug something that
was fixed while I wasn't looking ;)

It would be cool if there was a mailing list where commits
to the Xen BK repository were mailed out, to make it easier
for people to monitor what goes into the repository...

-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

RE: can't start domains

[James Harper]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 1481 bytes --]

it's polled rather than event driven (eg you have to go look at it), but this is a useful link for seeing what's new:
http://xen.bkbits.net:8080/xeno-unstable.bk

have a look at the followups to this message, try specifying the major/minor in hex form, or make sure the node exists in /dev. That was my problem in the end.

James







From: Rik van Riel
Sent: Mon 23/08/2004 3:10 PM
To: Ian Pratt
Cc: James Harper; xen-devel@lists.sourceforge.net; mike.wray@hpl.hp.com
Subject: Re: [Xen-devel] can't start domains


On Sun, 22 Aug 2004, Ian Pratt wrote:

> > So for some reason, I have a problem using any block device from xen!!!
> 
> I think we'd have noticed this if it were a general problem ;-)

I'm seeing the same problems with /dev/loop0 exported as
a VBD.

> Are you fully up to date with your repo? There have been a few
> duff ones in the last couple of weeks, but things seem stable
> now.

I'll upgrade the Xen utilities and xend tomorrow.  Hope
I didn't spend a few days trying to debug something that
was fixed while I wasn't looking ;)

It would be cool if there was a mailing list where commits
to the Xen BK repository were mailed out, to make it easier
for people to monitor what goes into the repository...

-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan

[-- Attachment #2: Type: text/html, Size: 2352 bytes --]

Re: can't start domains

[Ian Pratt]

> > Are you fully up to date with your repo? There have been a few
> > duff ones in the last couple of weeks, but things seem stable
> > now.
> 
> I'll upgrade the Xen utilities and xend tomorrow.  Hope
> I didn't spend a few days trying to debug something that
> was fixed while I wasn't looking ;)

I'd certainly recommend keeping all three components in sync if
you're using the unstable tree as there have been interface
changes. 
 
> It would be cool if there was a mailing list where commits
> to the Xen BK repository were mailed out, to make it easier
> for people to monitor what goes into the repository...

I guess we could set up a bk trigger that sent something to a new
email list...

Also, I think we should get back into the habit of using the bk
version tag to declare 'semi-stable' releases. It's apparently
been 11 weeks since we last did this; I've just added a tag to
the current repo as I'm pretty happy that it works for me.

I think we're very close to a 2.0 release. I guess the most
critical things are:
 * improve performance of 2.6 MUU interface
 * upgrade to 2.6.8.1

Wish list items would be:
 * finish cleanup of Xen foreign mapping grant table 
 * cleanup handling of multiple backend driver domains

Ian





-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

Re: can't start domains

[Ian Pratt]

> the hex solution will work nicely. Maybe xen needs its own set of major/minor numbers which xend can parse without having to look them up? eg /dev/xbd1

The trouble with using anything other than sd[a-z] hd[a-z] is
that it breaks a lot of scripts that various distros use. 

The way it currently works is that blkfront driver comes up early
in the boot process and uses a messaging channel to say to xend
"what block devices have I got?" If xend says 0801 (sda1) and
0343 (hdb3), then blkfront will grab block majors 08 and 03 and
create the appropriate disk structures.

One thing to watch out for is that if the blkfront driver has
been asked to grab major 08 then the scsi/sd layer won't be able
to get it, so that you won't then be able to have iscsi disks. If
you want to use iscsi as well, stick to using hd* (or xd*).

> I wonder how udev will work running in the xenU domain... do you think it would recognise the xen devices? i'll give it ago when I can afford some more testing time.

We currently don't have hooks in the blkfront driver to register
with devices udev. I expect these would be easy to add by someone
who knew udev.  We need to think through exactly how this will
interact with the move to passing device _names_ rather than
numbers.

Ian
 


-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

Re: can't start domains

[Rik van Riel]

On Mon, 23 Aug 2004, Ian Pratt wrote:

> I guess we could set up a bk trigger that sent something to a new
> email list...

I'd be happy to just pull the tree and send out such messages
myself.  I have access to the script that does the same thing
for the linux kernel.

As an aside, I got everything running fine here. I guess now
I'll have to push things until they break and send you guys
fixes and improvements ;)

-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

Re: can't start domains

[Ian Pratt]

> On Mon, 23 Aug 2004, Ian Pratt wrote:
> 
> > I guess we could set up a bk trigger that sent something to a new
> > email list...
> 
> I'd be happy to just pull the tree and send out such messages
> myself.  I have access to the script that does the same thing
> for the linux kernel.

Cool. I've requested a new sourceforge mailing list called
'xen-changelog' that the email could be fed too. 

The fact that people are reading the changelog might encourage us
to write more detailed entries too...
 
> As an aside, I got everything running fine here. 

Great -- I'm relieved to hear it.

> I guess now I'll have to push things until they break and send
> you guys fixes and improvements ;)

Here's hoping you have to try hard to break it ;-)

The 2.6 tree hasn't been given anywhere near as much testing as
2.4, but fingers crossed....

Ian



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

Re: can't start domains

[Rik van Riel]

On Mon, 23 Aug 2004, Ian Pratt wrote:

> Cool. I've requested a new sourceforge mailing list called
> 'xen-changelog' that the email could be fed too. 

I'll get on that list as soon as it exists.  The changelog
script is up and running.

> The fact that people are reading the changelog might encourage us
> to write more detailed entries too...

The changelogs mails can be generated automatically and
look like the mail below.  I'll run the changelog script
on the hour, so people will learn of new commits quickly.

---------- Forwarded message ----------
Date: Fri, 20 Aug 2004 18:19:24 +0000
From: BitKeeper Bot <riel@surriel.com>
Reply-To: Xen Development List <xen-devel@lists.sourceforge.net>
To: riel@surriel.com
Subject: Fix bad migration interaction with /lib/tls emulation code.

ChangeSet 1.1247.1.16, 2004/08/20 19:19:24+01:00, iap10@labyrinth.cl.cam.ac.uk

	Fix bad migration interaction with /lib/tls emulation code.



 reboot.c |    7 +++++++
 1 files changed, 7 insertions(+)


diff -Nru a/linux-2.6.7-xen-sparse/arch/xen/kernel/reboot.c b/linux-2.6.7-xen-sparse/arch/xen/kernel/reboot.c
--- a/linux-2.6.7-xen-sparse/arch/xen/kernel/reboot.c	2004-08-23 19:35:06 -04:00
+++ b/linux-2.6.7-xen-sparse/arch/xen/kernel/reboot.c	2004-08-23 19:35:06 -04:00
@@ -94,6 +94,13 @@

     HYPERVISOR_suspend(virt_to_machine(suspend_record) >> PAGE_SHIFT);

+    HYPERVISOR_vm_assist(VMASST_CMD_enable,
+			 VMASST_TYPE_4gb_segments);
+#ifdef CONFIG_XEN_WRITABLE_PAGETABLES
+    HYPERVISOR_vm_assist(VMASST_CMD_enable,
+			 VMASST_TYPE_writeable_pagetables);
+#endif
+
     shutting_down = -1;

     memcpy(&start_info, &suspend_record->resume_info, sizeof(start_info));



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

Re: can't start domains

[Ian Pratt]

> The changelogs mails can be generated automatically and
> look like the mail below.  I'll run the changelog script
> on the hour, so people will learn of new commits quickly.

I believe the xen-changelog@lists.sourceforge.net mailing list
now exists. I've set it up so that only <riel@surriel.com> can
post to it, but anyone can subscribe.

Thanks,
Ian


-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

Re: can't start domains

[Rik van Riel]

On Tue, 24 Aug 2004, Ian Pratt wrote:

> I believe the xen-changelog@lists.sourceforge.net mailing list
> now exists. I've set it up so that only <riel@surriel.com> can
> post to it, but anyone can subscribe.

It seems to be working fine.  I just got a few commits
back from the list.  I'll be running the script to
generate and send the commits on an hourly basis.

-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

Re: can't start domains

[Ian Pratt]

> On Tue, 24 Aug 2004, Ian Pratt wrote:
> 
> > I believe the xen-changelog@lists.sourceforge.net mailing list
> > now exists. I've set it up so that only <riel@surriel.com> can
> > post to it, but anyone can subscribe.
> 
> It seems to be working fine.  I just got a few commits
> back from the list.  I'll be running the script to
> generate and send the commits on an hourly basis.

The only problem seems to be that there's a 40KB message size
limit above which the list admin has to intervene. I can't figure
out how to bypass this.

I guess we don't have too many patches bigger than 40KB, so it's
probably not going to be a big deal for me to hit 'approve' every
so often. (Just my luck that there were two 40KB changesets
today). Perhaps I need to write a auto-approve-'bot ;-)

Rik: Thanks for setting up the auto-posting.

Ian

> 
> As list administrator, your authorization is requested for the
> following mailing list posting:
> 
>     List:    Xen-changelog@lists.sourceforge.net
>     From:    riel@surriel.com
>     Subject: Merge scramble.cl.cam.ac.uk:/auto/groups/xeno/BK/xeno.bk
>     Reason:  Message body is too big: 98245 bytes but there's a limit of 40 KB
> 
> At your convenience, visit:
> 
>     https://lists.sourceforge.net/lists/admindb/xen-changelog
>         
> to approve or deny the request.



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

RE: can't start domains

[James Harper]

[-- Attachment #1: Type: text/plain, Size: 2401 bytes --]

I'd question the usefulness of posting the entire patch. Speaking for myself, I'd be happy with a summary of the patch and a link to the bk changelog page (if possible) for anyone who wants to actually look at the patch. I would only subscribe to the changes list for notification of changes. Maybe a compromise could be to post the first 10k or so?

James



From: Ian Pratt
Sent: Wed 25/08/2004 8:18 AM
To: Rik van Riel
Cc: Ian Pratt; xen-devel@lists.sourceforge.net; Ian.Pratt@cl.cam.ac.uk
Subject: Re: [Xen-devel] can't start domains


> On Tue, 24 Aug 2004, Ian Pratt wrote:
> 
> > I believe the xen-changelog@lists.sourceforge.net mailing list
> > now exists. I've set it up so that only <riel@surriel.com> can
> > post to it, but anyone can subscribe.
> 
> It seems to be working fine.  I just got a few commits
> back from the list.  I'll be running the script to
> generate and send the commits on an hourly basis.

The only problem seems to be that there's a 40KB message size
limit above which the list admin has to intervene. I can't figure
out how to bypass this.

I guess we don't have too many patches bigger than 40KB, so it's
probably not going to be a big deal for me to hit 'approve' every
so often. (Just my luck that there were two 40KB changesets
today). Perhaps I need to write a auto-approve-'bot ;-)

Rik: Thanks for setting up the auto-posting.

Ian

> 
> As list administrator, your authorization is requested for the
> following mailing list posting:
> 
>     List:    Xen-changelog@lists.sourceforge.net
>     From:    riel@surriel.com
>     Subject: Merge scramble.cl.cam.ac.uk:/auto/groups/xeno/BK/xeno.bk
>     Reason:  Message body is too big: 98245 bytes but there's a limit of 40 KB
> 
> At your convenience, visit:
> 
>     https://lists.sourceforge.net/lists/admindb/xen-changelog
>         
> to approve or deny the request.



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

[-- Attachment #2: Type: text/html, Size: 2916 bytes --]

Re: can't start domains

[Niraj Tolia]

I agree with James on this one. Patch summaries are what I am
interested in. I can always use bk to get at the patch.

Niraj

----- Original Message -----
From: James Harper <jamesh@bendigoit.com.au>

 
I'd question the usefulness of posting the entire patch. Speaking for
myself, I'd be happy with a summary of the patch and a link to the bk
changelog page (if possible) for anyone who wants to actually look at
the patch. I would only subscribe to the changes list for notification
of changes. Maybe a compromise could be to post the first 10k or so?


-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285

RE: can't start domains

[Rik van Riel]

On Wed, 25 Aug 2004, James Harper wrote:

> I'd question the usefulness of posting the entire patch.

It makes it easy to reply to a particular code change and
ask questions about it on the mailing list.  It has proven
to be very effective for Linux kernel development.

> Maybe a compromise could be to post the first 10k or so?

Sounds like a good idea.  I guess I should figure a way to
set up a bk web interface here so I can post links for all
patches, plus the small patches to the list.

What do you think ?

Rik
-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285