Error: Received invalid response from Xend, twice.

Error: Received invalid response from Xend, twice.

[Andrew Thompson]

I sent this to xen-users yesterday and received no replies. Perhaps with a better title, it'll stand out to someone here:

----- Forwarded message from Andrew Thompson <andrewkt@aktzero.com> -----

Date: Mon, 3 Apr 2006 19:20:51 +0000
From: Andrew Thompson <andrewkt@aktzero.com>
To: xen-users@lists.xensource.com
Subject: [Xen-users] new box, won't boot domU

I just set up a new box that has been giving me trouble since I first cloned the xen repo to it.

I started off with "error: can't find a register in class 'BREG' while reloading 'asm'" errors that I fixed by changing my gcc profile(gentoo) to nopie. That introduced me to a '__guard' error that I fixed by changing my gcc profile to nopiesse. I will note for posterity that this was a new Gentoo install onto a old celeron with gcc 3.4.5-r1.

Now I've booted dom0 and have a file backed image that I'm trying to boot. I'm typing this from mutt running on a domU of a box sitting 4 foot away from the offending system, so I thought I'd be able to set it up without issue. (Suprise!)

My conversation looks like this:

backlash xen # xend start
backlash xen # xm create gdom1
Using config file "gdom1".
Error: Received invalid response from Xend, twice.
backlash xen # xend stop
backlash xen #

I truncated /var/log/xend.log beforehand and am including the entire dump of xend.log for the transcript above.

I think the below are the relevant bits:

[2006-04-03 13:26:22 xend.XendDomainInfo] INFO (XendDomainInfo:211) Recreating domain 8, UUID 39a4ef13-b02a-d201-dd37-081c6b99570a.
[2006-04-03 13:26:22 xend.XendDomainInfo] DEBUG (XendDomainInfo:199) XendDomainInfo.recreate({'paused': 1, 'cpu_time': 0L, 'ssidref': 0, 'handle': [57, 164, 239, 19, 176, 42, 210, 1, 221, 55, 8, 28, 107, 153, 87, 10], 'shutdown_reason': 0, 'dying': 0, 'dom': 9, 'mem_kb': 65536, 'maxmem_kb': 65536, 'max_vcpu_id': 0, 'crashed': 0, 'running': 0, 'shutdown': 0, 'online_vcpus': 1, 'blocked': 0})
[2006-04-03 13:26:22 xend.XendDomainInfo] INFO (XendDomainInfo:211) Recreating domain 9, UUID 39a4ef13-b02a-d201-dd37-081c6b99570a.
[2006-04-03 13:26:22 xend] ERROR (XendDomain:209) Failed to recreate information for domain 9.  Destroying it in the hope of recovery.
Traceback (most recent call last):
  File "/opt/xen-3.0-testing.hg/dist/install/usr/lib/python/xen/xend/XendDomain.py", line 203, in refresh
    self._add_domain(
  File "/opt/xen-3.0-testing.hg/dist/install/usr/lib/python/xen/xend/XendDomainInfo.py", line 235, in recreate
    vm = XendDomainInfo(xeninfo, domid, dompath, True, priv)
  File "/opt/xen-3.0-testing.hg/dist/install/usr/lib/python/xen/xend/XendDomainInfo.py", line 433, in __init__
    self.validateInfo()
  File "/opt/xen-3.0-testing.hg/dist/install/usr/lib/python/xen/xend/XendDomainInfo.py", line 564, in validateInfo
    self.check_name(self.info['name'])
  File "/opt/xen-3.0-testing.hg/dist/install/usr/lib/python/xen/xend/XendDomainInfo.py", line 1139, in check_name
    raise VmError("VM name '%s' is used in both domains %d and %d" %
VmError: VM name 'Domain-8' is used in both domains 9 and 8
[2006-04-03 13:26:22 xend] INFO (SrvServer:114) unix path=/var/lib/xend/xend-socket
[2006-04-03 13:26:22 xend.XendDomainInfo] DEBUG (XendDomainInfo:864) XendDomainInfo.handleShutdownWatch
[2006-04-03 13:26:22 xend.XendDomainInfo] DEBUG (XendDomainInfo:864) XendDomainInfo.handleShutdownWatch
[2006-04-03 13:26:34 xend] INFO (SrvDaemon:185) Xend stopped due to signal 15.

Afterwards(before the xend stop), I'm left with a paused Domain-8 in xm list.

Can I enable more debugging?

What should I attempt next?

Is it likely that this is related to the compiler profile I had to choose to get "make world" to complete(nopiesse)?

Any help appreciated.

-- 
Andrew Thompson
http://aktzero.com/


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

----- End forwarded message -----

-- 
Andrew Thompson
http://aktzero.com/

Re: Error: Received invalid response from Xend, twice.

[Ewan Mellor]

On Tue, Apr 04, 2006 at 03:51:47PM +0000, Andrew Thompson wrote:

> I sent this to xen-users yesterday and received no replies. Perhaps with a better title, it'll stand out to someone here:
> 
> ----- Forwarded message from Andrew Thompson <andrewkt@aktzero.com> -----
> 
> Date: Mon, 3 Apr 2006 19:20:51 +0000
> From: Andrew Thompson <andrewkt@aktzero.com>
> To: xen-users@lists.xensource.com
> Subject: [Xen-users] new box, won't boot domU
> 
> I just set up a new box that has been giving me trouble since I first cloned the xen repo to it.
> 
> I started off with "error: can't find a register in class 'BREG' while reloading 'asm'" errors that I fixed by changing my gcc profile(gentoo) to nopie. That introduced me to a '__guard' error that I fixed by changing my gcc profile to nopiesse. I will note for posterity that this was a new Gentoo install onto a old celeron with gcc 3.4.5-r1.
> 
> Now I've booted dom0 and have a file backed image that I'm trying to boot. I'm typing this from mutt running on a domU of a box sitting 4 foot away from the offending system, so I thought I'd be able to set it up without issue. (Suprise!)
> 
> My conversation looks like this:
> 
> backlash xen # xend start
> backlash xen # xm create gdom1
> Using config file "gdom1".
> Error: Received invalid response from Xend, twice.
> backlash xen # xend stop
> backlash xen #
> 
> I truncated /var/log/xend.log beforehand and am including the entire dump of xend.log for the transcript above.
> 
> I think the below are the relevant bits:
> 
> [Snip log]
> 
> Afterwards(before the xend stop), I'm left with a paused Domain-8 in xm list.
> 
> Can I enable more debugging?
> 
> What should I attempt next?
> 
> Is it likely that this is related to the compiler profile I had to choose to
> get "make world" to complete(nopiesse)?

The message "Error: Received invalid response from Xend, twice." means that
Xend has crashed whilst xm was talking to it, xm has retried, and it has
crashed again.  There is a known bug with this happening very intermittently,
but if you are seeing this permanently, I'd say yes, your strange compiler
options are almost certainly to blame.  You might try a full clean make with
those options turned on (you might have a mixed build by accident) but
otherwise, I think you're on your own, sorry.

Ewan.

Re: Error: Received invalid response from Xend, twice.

[Andrew Thompson]

On Tue, Apr 11, 2006 at 01:56:37PM +0100, Ewan Mellor wrote:
> The message "Error: Received invalid response from Xend, twice." means that
> Xend has crashed whilst xm was talking to it, xm has retried, and it has
> crashed again.  There is a known bug with this happening very intermittently,
> but if you are seeing this permanently, I'd say yes, your strange compiler
> options are almost certainly to blame.  You might try a full clean make with
> those options turned on (you might have a mixed build by accident) but
> otherwise, I think you're on your own, sorry.

Yesterday, I think, I discovered the line in my xend.log that confirms your diagnosis:

[2006-04-03 13:26:16 xend] CRITICAL (SrvDaemon:190) Xend died due to signal 9!  Restarting it.

I have tweaked the Gentoo CFLAGS and have an emerge -e world in progress. I will update the list afterwards.

-- 
Andrew Thompson
http://aktzero.com/