How to expand the file created by xenguest-install?

[Denis Forveille]

Does anyone has e detailed procedure to change the size (ie increase) of 
the file that is used as the "xvda" disk by a domU? This is the file 
created by the "xenguest-install" script (or the one created as in 
section 6.2 of the xen user's manual for the file-backed VBDs scenario).

Thanks in advance

[Ben]

Yeah, you use LVM and grow you logical volume size, then expand it in  
the guest domain. I'm not sure, but I suspect you'd have to restart  
your guest before you can expand it.

On Apr 23, 2006, at 5:18 PM, Denis Forveille wrote:
> Does anyone has e detailed procedure to change the size (ie  
> increase) of the file that is used as the "xvda" disk by a domU?  
> This is the file created by the "xenguest-install" script (or the  
> one created as in section 6.2 of the xen user's manual for the file- 
> backed VBDs scenario).
>
> Thanks in advance
>
> --
> Fedora-xen mailing list
> [e-mail]
> https://www.redhat.com/mailman/listinfo/fedora-xen

[Chris Hirsch]

Ben wrote:
> Yeah, you use LVM and grow you logical volume size, then expand it
> in the guest domain. I'm not sure, but I suspect you'd have to
> restart your guest before you can expand it.

But isn't the xvda just a file containing partitions and the like?
I've actually been wanting to do this myself WITH LVM. I'd LOVE to
have lvm extend the file itself.

I did find the trick extending an image:
dd if=/dev/zero  bs=1024 count=3500000 >> fedora.fc5.img
resize2fs -f fedora.fc5.img

and thats works great...The way I understand the xvda file is that
you'd have to copy out anything that is after the main image resize
the image (like above) and then append the old stuff back to the end
of the disk. I'd assume "stuff" here to mean something like swap.

Chris

[Ben]

Sorry, I should have been more clear. When you create the domU, you need to give 
it an LVM logical volume. As far as the domU is concerned, it's just a disk. But 
from the dom0 side, it's an easily expandible partition. After you grow it, use 
tools on the domU side to expand the filesystem.

There's really no point in trying to expand a swap partition because you can 
just make another one.

On Mon, 24 Apr 2006, Chris Hirsch wrote:
> Ben wrote:
> > Yeah, you use LVM and grow you logical volume size, then expand it
> > in the guest domain. I'm not sure, but I suspect you'd have to
> > restart your guest before you can expand it.
>
> But isn't the xvda just a file containing partitions and the like?
> I've actually been wanting to do this myself WITH LVM. I'd LOVE to
> have lvm extend the file itself.
>
> I did find the trick extending an image:
> dd if=/dev/zero  bs=1024 count=3500000 >> fedora.fc5.img
> resize2fs -f fedora.fc5.img
>
> and thats works great...The way I understand the xvda file is that
> you'd have to copy out anything that is after the main image resize
> the image (like above) and then append the old stuff back to the end
> of the disk. I'd assume "stuff" here to mean something like swap.
>
> Chris

[Jeremy Katz]

On Mon, 2006-04-24 at 20:29 -0400, Denis Forveille wrote:
> After that I'm stucked:
> I cant expand the filesystem for "/" from domU: "fdisk" gives the
>
> correct (new) size for "/dev/xvda", but I can't use parted or
> rezise2fs on "/dev/xvda1" as it is mounted...

Take a look at ext2online for online resizing of ext[23] filesystems
> Maybe, from dom0, a combination of "lomount" and "parted" or
>
> "resize2fs" or other tools could do the job but I don't know how to do
> it.

Take a look at kpartx (I think it's in the device-mapper-multipath
package for whatever reason, but that's neither here nor there) -- it'll
let you set up device mapper devices based on a partition table if you
want to do manipulation of guest filesystems from dom0

Jeremy

Multihomed networking

[Mark A Heilpern]

Is anyone using FC5 Xen with an FC5 guest, where the Dom0 has two ethernet devices (succesfully)?

My system has eth0 as a 192.168.2.x network, eth1 as my static Internet address, and iptables NAT between them. All real machines on my home network can get out just fine.

My guest VM, also a 192.168.2.x address, is able to reach other machines on my internal network without problem, but trying to access the internet doesn't work. At first I thought it was an iptables configuration issue on the Dom0 machine, but I don't think so anymore. As a test, I inserted at the top of every table a rule to emit a log message when any machine on my net tried to access a specific remote host; performing a wget command in the guest VM shows my outbound SYN, shows an incoming SYN+ACK from the remote host (so outbound masquarading must be working?) and an outbound ACK (so incoming demasquarading must be working?)... I never receive another packet from the remote host (until quite a while later, when I receive FIN). If I try the same exercise with a real host on my network, I get a normal data stream in/out.

Also, traceroute from the guest VM doesn't work exactly as I would expect. I would expect the results of one to be quite similar to those from my Dom0 machine, but consistantly I am missing responses along the path. For example:
> From dom0:

[root@heilpern ~]# traceroute www.yahoo.com
traceroute to www.yahoo.com (216.109.117.106), 30 hops max, 40 byte packets
 1  [my cable modem]
 2  10.106.32.1 (10.106.32.1)  8.279 ms   11.786 ms *
 3  * * *
 4  * * *
 5  * * *
 6  * * *
 7  ae-0-0.bbr2.Washington1.Level3.net (4.68.128.210)  45.855 ms * *
 8  ae-21-54.car1.Washington1.Level3.net (4.68.121.114)  35.593 ms ae-21-56.car1.Washington1.Level3.net (4.68.121.178)  35.382 ms   41.539 ms
 9  4.79.228.2 (4.79.228.2)  40.073 ms   39.931 ms *
10  * * *
11  * * *
12  * p21.www.dcn.yahoo.com (216.109.117.106)  35.227 ms *
> From the guest VM:

[root@testvm ~]# traceroute www.yahoo.com
traceroute to www.yahoo.com (216.109.117.106), 30 hops max, 40 byte packets
 1  192.168.2.1 (192.168.2.1)  0.000 ms   0.000 ms   0.000 ms
 2  [my cable modem]
 3  10.106.32.1 (10.106.32.1)  9.910 ms   13.919 ms *
 4  * * *
 5  * * *
 6  * * *
 7  * * *
 8  ae-0-0.bbr2.Washington1.Level3.net (4.68.128.210)  48.307 ms * *
 9  ae-11-51.car1.Washington1.Level3.net (4.68.121.18)  37.938 ms   42.377 ms ae-21-54.car1.Washington1.Level3.net (4.68.121.114)  43.048 ms
10  4.79.228.2 (4.79.228.2)  45.568 ms   43.918 ms *
11  * * *
12  * * *
13  * * *
14  * * *
15  * * *
16  * * *
17  * * *
18  * * *
19  * * *
20  * * *
21  * * *
22  * * *
23  * * *
24  * * *
25  * * *
26  * * *
27  * * *
28  * * *
29  * * *
30  * * *
[root@testvm ~]#

This is pretty repeatable.

Any thoughts?

Here are my Dom0 configurations:
[root@heilpern ~]# brctl show
bridge name     bridge id               STP enabled     interfaces
xenbr0          8000.feffffffffff       no              peth0
                                                        vif0.0
                                                        vif2.0
[root@heilpern ~]# ifconfig eth0
eth0      Link encap:Ethernet  HWaddr 00:10:DC:F2:70:17
          inet addr:192.168.2.1  Bcast:192.168.2.255  Mask:255.255.255.0
          inet6 addr: fe80::210:dcff:fef2:7017/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:733780 errors:0 dropped:0 overruns:0 frame:0
          TX packets:822036 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:87799449 (83.7 MiB)  TX bytes:207088442 (197.4 MiB)

[root@heilpern ~]# ifconfig eth1
eth1      Link encap:Ethernet  HWaddr 00:90:47:03:B8:63
          inet addr:[my IP]  Bcast:[my BCAST IP]  Mask:255.255.255.252
          inet6 addr: fe80::290:47ff:fe03:b863/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:829216 errors:0 dropped:0 overruns:0 frame:0
          TX packets:556902 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:169446616 (161.5 MiB)  TX bytes:85050877 (81.1 MiB)
          Interrupt:18

[root@heilpern ~]# iptables -t nat -L
Chain PREROUTING (policy ACCEPT)
target     prot opt source               destination

Chain POSTROUTING (policy ACCEPT)
target     prot opt source               destination
MASQUERADE  all  --  anywhere             anywhere

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination

(temporary wide-open rules):
[root@heilpern ~]# iptables -L
Chain INPUT (policy ACCEPT)
target     prot opt source               destination

Chain FORWARD (policy ACCEPT)
target     prot opt source               destination

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination

Chain INBOUND (0 references)
target     prot opt source               destination

Chain LOG_FILTER (0 references)
target     prot opt source               destination

Chain LSI (0 references)
target     prot opt source               destination

Chain LSO (0 references)
target     prot opt source               destination

Chain OUTBOUND (0 references)
target     prot opt source               destination

Thanks in advance...
Mark

Multihomed networking

[Ben]

Yes, but not with the default xen bridging scripts.

On Apr 24, 2006, at 6:27 PM, Mark A Heilpern wrote:
> Is anyone using FC5 Xen with an FC5 guest, where the Dom0 has two  
> ethernet devices (succesfully)?
>
> My system has eth0 as a 192.168.2.x network, eth1 as my static  
> Internet address, and iptables NAT between them. All real machines  
> on my home network can get out just fine.
>
> My guest VM, also a 192.168.2.x address, is able to reach other  
> machines on my internal network without problem, but trying to  
> access the internet doesn't work. At first I thought it was an  
> iptables configuration issue on the Dom0 machine, but I don't think  
> so anymore. As a test, I inserted at the top of every table a rule  
> to emit a log message when any machine on my net tried to access a  
> specific remote host; performing a wget command in the guest VM  
> shows my outbound SYN, shows an incoming SYN+ACK from the remote  
> host (so outbound masquarading must be working?) and an outbound  
> ACK (so incoming demasquarading must be working?)... I never  
> receive another packet from the remote host (until quite a while  
> later, when I receive FIN). If I try the same exercise with a real  
> host on my network, I get a normal data stream in/out.
>
> Also, traceroute from the guest VM doesn't work exactly as I would  
> expect. I would expect the results of one to be quite similar to  
> those from my Dom0 machine, but consistantly I am missing responses  
> along the path. For example:
> > From dom0:
>
> [root@heilpern ~]# traceroute www.yahoo.com
> traceroute to www.yahoo.com (216.109.117.106), 30 hops max, 40 byte  
> packets
>  1  [my cable modem]
>  2  10.106.32.1 (10.106.32.1)  8.279 ms   11.786 ms *
>  3  * * *
>  4  * * *
>  5  * * *
>  6  * * *
>  7  ae-0-0.bbr2.Washington1.Level3.net (4.68.128.210)  45.855 ms * *
>  8  ae-21-54.car1.Washington1.Level3.net (4.68.121.114)  35.593 ms  
> ae-21-56.car1.Washington1.Level3.net (4.68.121.178)  35.382 ms    
> 41.539 ms
>  9  4.79.228.2 (4.79.228.2)  40.073 ms   39.931 ms *
> 10  * * *
> 11  * * *
> 12  * p21.www.dcn.yahoo.com (216.109.117.106)  35.227 ms *
> > From the guest VM:
>
> [root@testvm ~]# traceroute www.yahoo.com
> traceroute to www.yahoo.com (216.109.117.106), 30 hops max, 40 byte  
> packets
>  1  192.168.2.1 (192.168.2.1)  0.000 ms   0.000 ms   0.000 ms
>  2  [my cable modem]
>  3  10.106.32.1 (10.106.32.1)  9.910 ms   13.919 ms *
>  4  * * *
>  5  * * *
>  6  * * *
>  7  * * *
>  8  ae-0-0.bbr2.Washington1.Level3.net (4.68.128.210)  48.307 ms * *
>  9  ae-11-51.car1.Washington1.Level3.net (4.68.121.18)  37.938 ms    
> 42.377 ms ae-21-54.car1.Washington1.Level3.net (4.68.121.114)   
> 43.048 ms
> 10  4.79.228.2 (4.79.228.2)  45.568 ms   43.918 ms *
> 11  * * *
> 12  * * *
> 13  * * *
> 14  * * *
> 15  * * *
> 16  * * *
> 17  * * *
> 18  * * *
> 19  * * *
> 20  * * *
> 21  * * *
> 22  * * *
> 23  * * *
> 24  * * *
> 25  * * *
> 26  * * *
> 27  * * *
> 28  * * *
> 29  * * *
> 30  * * *
> [root@testvm ~]#
>
> This is pretty repeatable.
>
> Any thoughts?
>
> Here are my Dom0 configurations:
> [root@heilpern ~]# brctl show
> bridge name     bridge id               STP enabled     interfaces
> xenbr0          8000.feffffffffff       no              peth0
>                                                         vif0.0
>                                                         vif2.0
> [root@heilpern ~]# ifconfig eth0
> eth0      Link encap:Ethernet  HWaddr 00:10:DC:F2:70:17
>           inet addr:192.168.2.1  Bcast:192.168.2.255  Mask: 
> 255.255.255.0
>           inet6 addr: fe80::210:dcff:fef2:7017/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:733780 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:822036 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:0
>           RX bytes:87799449 (83.7 MiB)  TX bytes:207088442 (197.4 MiB)
>
> [root@heilpern ~]# ifconfig eth1
> eth1      Link encap:Ethernet  HWaddr 00:90:47:03:B8:63
>           inet addr:[my IP]  Bcast:[my BCAST IP]  Mask:255.255.255.252
>           inet6 addr: fe80::290:47ff:fe03:b863/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:829216 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:556902 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:1000
>           RX bytes:169446616 (161.5 MiB)  TX bytes:85050877 (81.1 MiB)
>           Interrupt:18
>
> [root@heilpern ~]# iptables -t nat -L
> Chain PREROUTING (policy ACCEPT)
> target     prot opt source               destination
>
> Chain POSTROUTING (policy ACCEPT)
> target     prot opt source               destination
> MASQUERADE  all  --  anywhere             anywhere
>
> Chain OUTPUT (policy ACCEPT)
> target     prot opt source               destination
>
> (temporary wide-open rules):
> [root@heilpern ~]# iptables -L
> Chain INPUT (policy ACCEPT)
> target     prot opt source               destination
>
> Chain FORWARD (policy ACCEPT)
> target     prot opt source               destination
>
> Chain OUTPUT (policy ACCEPT)
> target     prot opt source               destination
>
> Chain INBOUND (0 references)
> target     prot opt source               destination
>
> Chain LOG_FILTER (0 references)
> target     prot opt source               destination
>
> Chain LSI (0 references)
> target     prot opt source               destination
>
> Chain LSO (0 references)
> target     prot opt source               destination
>
> Chain OUTBOUND (0 references)
> target     prot opt source               destination
>
> Thanks in advance...
> Mark
>
> --
> Fedora-xen mailing list
> [e-mail]
> https://www.redhat.com/mailman/listinfo/fedora-xen

[Denis Forveille]

Thanks, but I still miss how to grow the partition inside the disk 
inside the lvg (lvg -> /dev/xvda -> /dev/xvda1)

To summarize, the steps should be (from dom0):
1) grow the lvg used for domU (=/dev/xvda from the domU point of view)
    -> lvextend -L +1024 /dev/vg_domU/lv_FC55
2) grow the partition used by "/" in the lvg (/dev/xvda1) with parted
    -> parted /dev/mapper/vg_domU-lv_FC55
    -> resize 1 32kB 9664MB (This does not work, see below)
3) resize the fs with resize2fs (from dom0) or with ext2online from domU 
(thanks to Jeremy)
    -> resize2fs /dev/mapper/vg_domU-lv_FC55

Here's the output from parted:
   Using /dev/mapper/vg_domU-lv_FC55
   (parted) print
   Disk geometry for /dev/mapper/vg_domU-lv_FC55: 0kB - 9664MB
   Disk label type: msdos
   Number  Start   End     Size    Type      File system  Flags
   1       32kB    8587MB  8587MB  primary   ext3         boot
   (parted) resize 1
   Start?  [32kB]? 32kB
   End?  [8587MB]? 9660MB
   Error: File system has an incompatible feature enabled.

Is there a tool other than parted I could use to resize the partition 
/dev/xvda1 inside the lvg?

Jeremy Katz wrote:
> On Mon, 2006-04-24 at 20:29 -0400, Denis Forveille wrote:
> > After that I'm stucked:
> > I cant expand the filesystem for "/" from domU: "fdisk" gives the
> >
> > correct (new) size for "/dev/xvda", but I can't use parted or
> > rezise2fs on "/dev/xvda1" as it is mounted...
>
> Take a look at ext2online for online resizing of ext[23] filesystems
> > Maybe, from dom0, a combination of "lomount" and "parted" or
> >
> > "resize2fs" or other tools could do the job but I don't know how to do
> > it.
>
> Take a look at kpartx (I think it's in the device-mapper-multipath
> package for whatever reason, but that's neither here nor there) -- it'll
> let you set up device mapper devices based on a partition table if you
> want to do manipulation of guest filesystems from dom0
>
> Jeremy
>
> --
> Fedora-xen mailing list
> [e-mail]
> https://www.redhat.com/mailman/listinfo/fedora-xen

[Lamont R. Peterson]

On Monday 24 April 2006 06:41pm, Jeremy Katz wrote:
> On Mon, 2006-04-24 at 20:29 -0400, Denis Forveille wrote:
> > After that I'm stucked:
> > I cant expand the filesystem for "/" from domU: "fdisk" gives the
> >
> > correct (new) size for "/dev/xvda", but I can't use parted or
> > rezise2fs on "/dev/xvda1" as it is mounted...
>
> Take a look at ext2online for online resizing of ext[23] filesystems

Is ext2online supposed to work for ext2 filesystems?  I have read that it only 
worked for ext3.  I've tried it with ext2 (created using mke2fs on RHEL4) and 
it never worked.

[snip]

[Lamont R. Peterson]

On Monday 24 April 2006 08:45pm, Denis Forveille wrote:
> Thanks, but I still miss how to grow the partition inside the disk
> inside the lvg (lvg -> /dev/xvda -> /dev/xvda1)

I don't have partitions inside of the LVs in dom0.  It works great for all 
distros (SUSE 9.x/10.0, SLES8/9, RHEL3/4, CentOS3/4, FC2/3/4/5, Gentoo, 
Ubuntu).  Just make the LV the partition for the domU guests.  Then, you 
don't have to worry about a partition table from either dom0 or domU and the 
guests don't care about it either.

[snip]

[Jeremy Katz]

On Tue, 2006-04-25 at 11:14 -0600, Lamont R. Peterson wrote:
> On Monday 24 April 2006 06:41pm, Jeremy Katz wrote:
> > On Mon, 2006-04-24 at 20:29 -0400, Denis Forveille wrote:
> > > After that I'm stucked:
> > > I cant expand the filesystem for "/" from domU: "fdisk" gives the
> > >
> > > correct (new) size for "/dev/xvda", but I can't use parted or
> > > rezise2fs on "/dev/xvda1" as it is mounted...
> >
> > Take a look at ext2online for online resizing of ext[23] filesystems
>
> Is ext2online supposed to work for ext2 filesystems?  I have read that it only 
> worked for ext3.  I've tried it with ext2 (created using mke2fs on RHEL4) and 
> it never worked.

I thought the support went in, but I could be mistaken.  As it turns
out, I never create ext2 filesystems anymore :)

Jeremy

Expanding a VBD in FC5

[Kwan Lowe]

Okay... I think I've figured out how to do this..

Problem:  I created a volume backed DomU on a 3G logical volume. The DomU itself
used LVM to manage disks. I needed to expand the disk (i.e., the LVM on the Dom0).

The previous workaround was to add another LV in the Dom0 and assign it to the DomU.
This created a second xvdb in the domU. I used standard LVM tools to create a PV and
assign it to the original volume group.  E.g.:

My xm-fc5-005 initially contained:
  name = "xm-fc5-005"
  memory = "256"
  disk = [ 'phy:/dev/rootvg/xm-fc4-005_lv,xvda,w']
  vif = [ 'mac=00:16:3e:20:48:ae' ]
  bootloader="/usr/bin/pygrub"

I added another disk by creating a new LV in dom0 and assigning with:

disk = [ 'phy:/dev/rootvg/xm-fc4-005_lv,xvda,w',\
          'phy:/dev/rootvg/xm-fc5-005_lv2,xvdb,w' ]

The above works fine, and is probably no less efficient than the method that
follows, but it felt kind of kludgey to add space that way. The second method is to
lvresize the original VBD and use pvresize in the domU:

lvresize -L +1G /dev/rootvg/xm-fc4-005_lv

Then in the domU:
First, determine the PV device by running pvdisplay and getting the PV Name. In this
case, xvda2. Next resize it:

pvresize --setphysicalvolumesize 4G /dev/xvda2

pvresize appears to only be available in FC5... alas.

If you're not running LVMs inside the domU, you'd use parted (or qparted) to resize.