Automatic loopback dev with 'whole disk' stored in file

Automatic loopback dev with 'whole disk' stored in file

[Stephen Childs]

I am (finally) porting my setup code to use the automatic loopback 
disk='file:/blah' stuff rather than setting up the loopback devices myself. 
This doesn't seem to work when the file is backing a whole 'device' rather 
than just a partition hosting a filesystem? (i.e. the file actually contains 
partitions which then hold filesystems)

I get the following error on the command line:

# xm create name=cagnode89 disk=file:/var/xen-grid/cagnode89/fs/vdisk,hda1,w 
ip=134.226.53.153 root=/dev/hda2
Using config file "/etc/xen/xmdefconfig".
Error: Error creating domain: [Errno 4] Interrupted system call


and this in xend.log:

[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:720) init_domain> Created 
domain=9 name=cagnode89 memory=200
[2004-10-28 12:39:52 xend] INFO (console:94) Created console id=12 domain=9 
port=9609
[2004-10-28 12:39:52 xend] DEBUG (blkif:115) Connecting blkif 
<BlkifBackendInterface 9 0>
[2004-10-28 12:39:52 xend] INFO (XendRoot:91) EVENT> xend.console.create 
[12, 9, 9609]
[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:1082) Creating vbd dom=9 
uname=file:/var/xen-grid/cagnode89/fs/vdisk
[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:674) Destroying vbds for 
domain 9
[2004-10-28 12:39:52 xend] DEBUG (blkif:501) Destroying blkif domain=9
[2004-10-28 12:39:52 xend] DEBUG (blkif:357) Destroying vbd domain=9 idx=0
[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:634) Closing console, domain 9
[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:622) Closing channel to 
domain 9

If I do the losetup manually and then use phy:loop0 then it still works fine.

Maybe I should move over to using one file per partition as that seems to be 
what is being tested most. Is it just me who thinks it's quite nice to have 
the whole disk, partitions and all, stored in a single file?

Stephen
-- 
Dr. Stephen Childs,
Research Fellow, EGEE Project,    phone:                    +353-1-6081797
Computer Architecture Group,      email:        Stephen.Childs @ cs.tcd.ie
Trinity College Dublin, Ireland   web: http://www.cs.tcd.ie/Stephen.Childs


-------------------------------------------------------
This Newsletter Sponsored by: Macrovision 
For reliable Linux application installations, use the industry's leading
setup authoring tool, InstallShield X. Learn more and evaluate 
today. http://clk.atdmt.com/MSI/go/ins0030000001msi/direct/01/

Automatic loopback dev with 'whole disk' stored in file

[Stephen Childs]

I am (finally) porting my setup code to use the automatic loopback 
disk='file:/blah' stuff rather than setting up the loopback devices myself. 
This doesn't seem to work when the file is backing a whole 'device' rather 
than just a partition hosting a filesystem? (i.e. the file actually contains 
partitions which then hold filesystems)

I get the following error on the command line:

# xm create name=cagnode89 disk=file:/var/xen-grid/cagnode89/fs/vdisk,hda1,w 
ip=134.226.53.153 root=/dev/hda2
Using config file "/etc/xen/xmdefconfig".
Error: Error creating domain: [Errno 4] Interrupted system call


and this in xend.log:

[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:720) init_domain> Created 
domain=9 name=cagnode89 memory=200
[2004-10-28 12:39:52 xend] INFO (console:94) Created console id=12 domain=9 
port=9609
[2004-10-28 12:39:52 xend] DEBUG (blkif:115) Connecting blkif 
<BlkifBackendInterface 9 0>
[2004-10-28 12:39:52 xend] INFO (XendRoot:91) EVENT> xend.console.create 
[12, 9, 9609]
[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:1082) Creating vbd dom=9 
uname=file:/var/xen-grid/cagnode89/fs/vdisk
[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:674) Destroying vbds for 
domain 9
[2004-10-28 12:39:52 xend] DEBUG (blkif:501) Destroying blkif domain=9
[2004-10-28 12:39:52 xend] DEBUG (blkif:357) Destroying vbd domain=9 idx=0
[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:634) Closing console, domain 9
[2004-10-28 12:39:52 xend] DEBUG (XendDomainInfo:622) Closing channel to 
domain 9

If I do the losetup manually and then use phy:loop0 then it still works fine.

Maybe I should move over to using one file per partition as that seems to be 
what is being tested most. Is it just me who thinks it's quite nice to have 
the whole disk, partitions and all, stored in a single file?

Stephen


-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click

Re: Automatic loopback dev with 'whole disk' stored in file

[Ian Pratt]

> I am (finally) porting my setup code to use the automatic loopback 
> disk='file:/blah' stuff rather than setting up the loopback devices myself. 
> This doesn't seem to work when the file is backing a whole 'device' rather 
> than just a partition hosting a filesystem? (i.e. the file actually contains 
> partitions which then hold filesystems)

I don't believe the standard loop back device supports whole
device. Are you using the NASA enhanced loop back device patch?

I guess you'd need to use this patch in domain 0 to populate the
file in the first place. The other domains won't care.

However, It's not immediately obvious to me why the standard
script that handles setting up loopback devices cares whether
you're exporting partitions or a whole disk.

Any reason not to have a file per partition? (Unless you want to
let your domains partition their 'disk' however they see fit.)

Ian




-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click