Bug 645237

Fix built into xen-3.0.3-118.el5

Version-Release number of selected component (if applicable):
libvirt-0.8.2-8.el5
kernel-2.6.18-227.el5
xen-3.0.3-117.el5

Steps to Reproduce:
1. create a vm with attached xml file
   # virsh define $test.xml
   # virsh start $vm_name
2. try to attach an unexisting iso to vm 
# virsh attach-disk hello /var/lib/libvirt/images/test.iso hdc --driver xen
--subdriver raw --type cdrom --mode readonly --sourcetype file
Disk attached successfully

3. dump guest xml to check whether the cdrom is attached.
   # virsh dumpxml hello |grep cdrom -A 5 -B 5 
    <disk type='file' device='disk'>
      <driver name='file'/>
      <source file='/var/lib/xen/images/RHEL-Server-5.5-32-hvm.raw'/>
      <target dev='hda' bus='ide'/>
    </disk>
    <disk type='block' device='cdrom'>
      <driver name='xen'/>
      <source dev='/var/lib/libvirt/images/test.iso'/>
      <target dev='hdc' bus='ide'/>
      <readonly/>


Actual results:
could attach unexisting cdrom to guest without error message

Version-Release number of selected component (if applicable):
libvirt-0.8.2-8.el5
kernel-2.6.18-227.el5
xen-3.0.3-118.el5

Actual steps：
1. create a vm with attached xml file
   # virsh define $test.xml
   # virsh start $vm_name

2. try to attach an unexisting iso to vm。
  # virsh attach-disk hello /var/lib/libvirt/images/test.iso hdc --driver xen --subdriver raw --type cdrom --mode readonly --sourcetype file

error: Failed to attach disk
error: Requested operation is not valid: Xm driver only supports modifying persistent config

3. dump guest xml to check whether the cdrom is attached.
   # virsh dumpxml hello |grep cdrom -A 5 -B 5 
   <disk type='file' device='disk'>
      <driver name='file'/>
      <source file='/var/lib/xen/images/RHEL-Server-5.5-32-hvm.raw'/>
      <target dev='hda' bus='ide'/>
    </disk>
    <disk type='file' device='cdrom'>
      <target dev='hdc' bus='ide'/>
      <readonly/>
    </disk>
    <interface type='bridge'>
      <mac address='00:16:36:27:22:e3'/>

Actual results:
1. guest prompts error msgs.

In comment#6, with xen-118, we can see cdrom through "xm block-list"

# xm block-list 2

 Vdev  BE handle state evt-ch ring-ref BE-path
768    0    0     6      4      8     /local/domain/0/backend/vbd/2/768  
5632    0    0     3      5      9     /local/domain/0/backend/vbd/2/5632 

But if we use "xm block-attach",  we CANNOT see cdrom through "xm block-list".

1). # xm block-list 5
Vdev  BE handle state evt-ch ring-ref BE-path
768    0    0     6      4      8     /local/domain/0/backend/vbd/5/768


2). # xm block-attach 5 file:/var/lib/libvirt/images/test.iso hdc:cdrom r
Error: Device 5632 (vbd) could not be connected. /etc/xen/scripts/block failed;
error detected.
Usage: xm block-attach <Domain> <BackDev> <FrontDev> <Mode>

Create a new virtual block device.

3). # xm block-list 5
Vdev  BE handle state evt-ch ring-ref BE-path
768    0    0     6      4      8     /local/domain/0/backend/vbd/5/768  

Should I change it to verified?

Hi Yang,

if attach fail, there's no cdrom in guest so block-list is correct. As comment #6  shows, error is reported so the bz can be set to verified.

During "virsh disk-attached". we can find a cdrom with "xm block-list". But after  "virsh disk-attached",  an error info occurs and cdrom is removed.

Considering the patch just modify the reconfigureDevice() method, I don't think this bug should be verified with 'xm block-attach', which has nothing to do with reconfigureDevice(). Looking into xm codes, I find that 'xm block-configure' is the right one to verify this bug, which triggers reconfigureDevice() to change cdrom device. Bellow is my step to verify this bug(patch?):

(1) Create a HVM guest, with cdrom device configured in config file like:

disk = ['file:/xen-autotest/xen-autotest/client/tests/xen/images/RHEL-Server-5.5-64-hvm.raw,hda,w', 'file:/root/sol-10-u9-ga-x86-dvd.iso,hdc:cdrom,r']

(2) Reconfigure the cdrom device via xm block-configure command:

#xm block-configure vm1 file:/root/bot.iso hdd:cdrom r

You could use a non-existed ISO as the back-end device. However, there is misspelling in the patch:

if not os.access(new_back['params'],os.R_OK):
    raise VmError("Can't read disk file %" % new_back['params'])

should be:

if not os.access(new_back['params'],os.R_OK):
    raise VmError("Can't read disk file %s" % new_back['params'])

Thus we could only get "ValueError: incomplete format" output by now from the command. I have to reopen this bug so that the fix could be fixed in the next version.

(In reply to comment #11)
> Considering the patch just modify the reconfigureDevice() method, I don't think
> this bug should be verified with 'xm block-attach', which has nothing to do
> with reconfigureDevice(). Looking into xm codes, I find that 'xm
> block-configure' is the right one to verify this bug, which triggers
> reconfigureDevice() to change cdrom device. Bellow is my step to verify this
> bug(patch?):
> 
> (1) Create a HVM guest, with cdrom device configured in config file like:
> 
> disk =
> ['file:/xen-autotest/xen-autotest/client/tests/xen/images/RHEL-Server-5.5-64-hvm.raw,hda,w',
> 'file:/root/sol-10-u9-ga-x86-dvd.iso,hdc:cdrom,r']
> 
> (2) Reconfigure the cdrom device via xm block-configure command:
> 
> #xm block-configure vm1 file:/root/bot.iso hdd:cdrom r

should be:

#xm block-configure vm1 file:/root/bot.iso hdc:cdrom r

Sorry for inconvenience :)


> 
> You could use a non-existed ISO as the back-end device. However, there is
> misspelling in the patch:
> 
> if not os.access(new_back['params'],os.R_OK):
>     raise VmError("Can't read disk file %" % new_back['params'])
> 
> should be:
> 
> if not os.access(new_back['params'],os.R_OK):
>     raise VmError("Can't read disk file %s" % new_back['params'])
> 
> Thus we could only get "ValueError: incomplete format" output by now from the
> command. I have to reopen this bug so that the fix could be fixed in the next
> version.

Fix built into xen-3.0.3-120.el5

Try to verify this bug:

Actual steps:
same with comment#11

Actual Results:
1). With xen-117 and  kernel-xen-233, host doesn't print any error message.
2). With xen-118 and  kernel-xen-233, host prints an error message:"incomplete format".
3). With xen-118 and  kernel-xen-233, host prints an error message:
"Error: Can't read disk file /root/bot.iso
Usage: xm block-configure <Domain> <BackDev> <FrontDev> <Mode> [BackDomId]

Change block device configuration
"

With results, I change it to verified.

Sorry, the result 3).
3).  With xen-120 and  kernel-xen-233.



(In reply to comment #15)
> Try to verify this bug:
> 
> Actual steps:
> same with comment#11
> 
> Actual Results:
> 1). With xen-117 and  kernel-xen-233, host doesn't print any error message.
> 2). With xen-118 and  kernel-xen-233, host prints an error message:"incomplete
> format".
> 3). With xen-118 and  kernel-xen-233, host prints an error message:
> "Error: Can't read disk file /root/bot.iso
> Usage: xm block-configure <Domain> <BackDev> <FrontDev> <Mode> [BackDomId]
> 
> Change block device configuration
> "
> 
> With results, I change it to verified.

An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHBA-2011-0031.html