Problem booting jailtime images
by Murat Işık
Hello,
After installing a fresh fc5, I followed
fedoraproject.org/wiki/FedoraXenQuickstartFC5 up until xenguest-install.py
part. After that point for a quick try I downloaded both fc4 and debian
prebuilt images from jailtime.org . However both images, as I boot them
with xm create -c {related one}.cfg , crash with this error:
VFS: Cannot open root device "sda1" or unknown-block(0,0)
Please append a correct "root=" boot option
Kernel panic - not syncing: VFS: Unable to mount root fs on
unknown-block(0,0)
I did extensive googling and mounted the images, cheking their fstab and
so on. Everything looks fine. But the problem is unresolved.
Does anybody have any experiences with jailtime images? any suggestions?
BTW my host fc5 has LVM partitioning. Very unlikely but can this have an
effect?
Have a nice day.
--
Murat Isik
www.bilgiguvenlik.com
18 years, 3 months
Xen 3.0 on FC5 problems
by Filipe Miranda
Hello there,
I'm new to this xen technology and I'm trying it on FC5 but I'm having a
hardtime trying to setup a guest xen machine.
Here is what I did and all the error messages:
The directory /nfs/fc5 has the following content:
[root@fc5 fc5]# ls /nfs/fc5/
Fedora figs images isolinux repodata stylesheet-images
I folllowed the steps described on
http://fedoraproject.org/wiki/FedoraXenQuickstartFC5
[root@fc5 fc5]# /usr/sbin/xenguest-install.py
What is the name of your virtual machine? fc5
How much RAM should be allocated (in megabytes)? 256
What would you like to use as the disk (path)? /xen/fc5
What is the install location? nfs:fc5:/nfs/fc5
Starting install...
Using config file "/etc/xen/fc5".
Error: Device 51712 (vbd) could not be connected. Backend device not found.
If your install has exited, you can restart your guest by running
'xm create -c fc5'. Otherwise, you can reconnect to the console
by running 'xm console fc5'
Any suggestions?
--
Att.
---
Filipe T Miranda
18 years, 3 months
VMX guest machine issue
by sxu1902
Hi,
I am testing Xen3 on FC5. My FC5 is x86_64. My box is using Pentium D 920 on Asus P5LD2-VM motherboard with 2G RAM.
After installation, "xm list" shows domain0 running ok.
I use:
dd if=/dev/zero of=/hdd_mnt/sda5/g_img/gl112/gl112.img bs=1M count=1 seek=10240
to create iamge file.
And the paremeter file look like this:
# Automatically generated xen config file
name = "gl112"
builder = "hvm"
memory = "256"
acpi = 0
apic = 0
cdrom = '/hdd_mnt/sda5/dvd/CentOS-4.3-i386-binDVD.iso'
#cdrom = '/dev/cdrom'
boot = 'd'
disk = [ 'file:/hdd_mnt/sda5/g_img/gl112/gl112.img,ioemu:hda,w' ]
vif = [ 'type=ioemu,bridge=xenbr0,mac=00:17:3e:59:78:30' ]
on_reboot = 'restart'
on_crash = 'restart'
#on_poweroff = 'destroy'
kernel = '/usr/lib/xen/boot/hvmloader'
device_model = '/usr/lib64/xen/bin/qemu-dm'
sdl = 1 # use SDL for graphics
vnc = 0 # use VNC for graphics
vncviewer = 0 # spawn vncviewer by default
nographic = 0 # don't use graphics
serial = 'pty' # enable serial console
I tried different linux distributions as VMX guest. It appears to me the guest machine always reboot itself when loading the linux from CD images or CDs. Sometimes it even freezes domain0. Occasionally it boots into the installation CD (after reboot itself once) and tells me it can't find hard drive, so it can't proceed with the installation...
Anybody tested VMX? Any similar problems? Any suggestions?
Sam Xu
18 years, 3 months
RE: [Fedora-xen] VMX guest machine issue
by You, Yongkang
Hi Sam,
FC5 release has some known issue to play with VMX in x86_64, e.g. destroying IA32 VMX will hang Xen0 and creating x86_64 VMX will reboot Xen0. This bug will be fixed in the FC5 update. I am not sure if it is there now. Seemed not?
A way is worthy to have a try, if you don’t want to wait for FC5 update to test. Get the Xen source RPM package from FC5 and only re-make xen (go into xen folder to do make, you don’t need to remake xen0 and xenU). After making finishes, copy xen.gz to /boot and replace the old hypervisor kernel. Reboot~.
I didn’t find any error in your config file. But the error info “it can't find hard drive“ is very strange. Did you try the FC5 or FC4 ISO to install? Have a look at /var/log/xend.log to see if the disk option (hda) has been sent to qemu, such like:
/usr/lib64/xen/bin/qemu-dm', '-d', '1’, '-m', '256', '-boot', 'c', '-serial', 'pty', '-vcpus', '1', '-domain-name', 'ExampleHVMDomain', '-hda', '/xxx/xxx.img', '-hdb', '/xxx/yyy.img', '-macaddr', '00:16:3e:00:00:11', '-bridge', 'xenbr0', '-nics', '1'
BTW, you can find a lot of OS installed disk image in http://free.oszoo.org/ .
Best Regards,
Yongkang (Kangkang) 永康
________________________________
From: fedora-xen-bounces(a)redhat.com [mailto:fedora-xen-bounces@redhat.com] On Behalf Of sxu1902
Sent: 2006年3月29日 23:08
To: fedora-xen(a)redhat.com
Subject: [Fedora-xen] VMX guest machine issue
Hi,
I am testing Xen3 on FC5. My FC5 is x86_64. My box is using Pentium D 920 on Asus P5LD2-VM motherboard with 2G RAM.
After installation, "xm list" shows domain0 running ok.
I use:
dd if=/dev/zero of=/hdd_mnt/sda5/g_img/gl112/gl112.img bs=1M count=1 seek=10240
to create iamge file.
And the paremeter file look like this:
# Automatically generated xen config file
name = "gl112"
builder = "hvm"
memory = "256"
acpi = 0
apic = 0
cdrom = '/hdd_mnt/sda5/dvd/CentOS-4.3-i386-binDVD.iso'
#cdrom = '/dev/cdrom'
boot = 'd'
disk = [ 'file:/hdd_mnt/sda5/g_img/gl112/gl112.img,ioemu:hda,w' ]
vif = [ 'type=ioemu,bridge=xenbr0,mac=00:17:3e:59:78:30' ]
on_reboot = 'restart'
on_crash = 'restart'
#on_poweroff = 'destroy'
kernel = '/usr/lib/xen/boot/hvmloader'
device_model = '/usr/lib64/xen/bin/qemu-dm'
sdl = 1 # use SDL for graphics
vnc = 0 # use VNC for graphics
vncviewer = 0 # spawn vncviewer by default
nographic = 0 # don't use graphics
serial = 'pty' # enable serial console
I tried different linux distributions as VMX guest. It appears to me the guest machine always reboot itself when loading the linux from CD images or CDs. Sometimes it even freezes domain0. Occasionally it boots into the installation CD (after reboot itself once) and tells me it can't find hard drive, so it can't proceed with the installation...
Anybody tested VMX? Any similar problems? Any suggestions?
Sam Xu
你 知 道 2006 年 最 火 爆 的 电 子 邮 箱 吗 ?
多 达 1.1 亿 中 国 人 选 择 网 易 2000 兆 免 费 邮 箱 , 抢 注 再 得 280 兆 超 大 网 盘 <http://www.126.com/help/126fab.htm>
18 years, 3 months
Re: [Fedora-xen] Is guest-os only FC5 ?
by Roy-Magne Mo
on den 29.03.2006 klokka 14:26 (-0500) skreiv Robert Story:
> On Wed, 29 Mar 2006 17:16:47 +0200 Roy-Magne wrote:
> RMM> You can only install distros which have a native XEN guest kernel, AFAIK
> RMM> this is only FC5 and rawhide at the moment.
>
> That may be a restriction imposed by the FC5 anaconda installer, but I don't
> think it's true in general.
>
> Although I'm using the FC5 domU kernel, I've got 2 domUs running CentOS 4.3. A
> little more work to install, but they are working fine.
I don't mean that FC5 is the only distribution that could run under XEN
in Fedora, but it's problably one of the few that can be installed
unmodified.
--
Roy-Magne Mo <rmo(a)sunnmore.net>
18 years, 3 months
Re: [Fedora-xen] Is guest-os only FC5 ?
by tournesol
I think it is not slash problem
let me show you detail
# xenguest-install.py
What is the name of your virtual machine? fc4
How much RAM should be allocated (in megabytes)? 260
What would you like to use as the disk (path)? /opt/fc4/
What is the install location?
ftp://ftp.riken.jp/Linux/fedora/core/4/i386/os/
Invalid URL location given
Otter wrote:
>You are still missing the slash here:
>
>What would you like to use as the disk (path)? /opt/fc4
>
>use
>What would you like to use as the disk (path)? /opt/fc4/
>
>
>
>
18 years, 3 months
Automated installations via kickstart
by Kwan Lowe
Whoohoo! Finally got xen/kickstart to do a complete hands-off installation. The
goal is to be able to provision an OS for a customer (internal/external) with as
little interaction as possible, giving me more time to drink my coffee :D. The
kickstart piece makes it as easy/easier than AIX NIM installations...
I used the anaconda_ks.cfg from a previously installed domU and made a few changes.
To use it with xenguest-install.py you need to specify the kickstart file with the
"-x" option:
xenguest-install.py -n xm-fc4-004 \
-f /dev/rootvg/xm-fc4-004_lv \
-r 256 -l ftp://192.168.8.21/pub \
-x ks=ftp://192.168.8.21/xm-fc4-004_ks.cfg
The kickstart file follows.
Some notes:
If you start with an anaconda generated file you need to add the keyboard directive.
For some reason it does not save this information (bug??). The zerombr directive
will also destroy the existing LVM volume so use with care. The rootpw directive
needs to be changed in the example. You'd probably want to change the bootproto
option also. I'm hand-editing the IP info because I already have DNS entries.
# Kickstart file automatically generated by anaconda.
install
url --url ftp://192.168.8.21//pub
lang en_US.UTF-8
keyboard us
network --device eth0 --bootproto static --ip 192.168.8.64 --netmask 255.255.255.0
--gateway 192.168.8.1 --nameserver 192.168.8
rootpw --iscrypted $1$ma1gooDL$myPaSsw0rdhashwh000h000
firewall --enabled --port=22:tcp
authconfig --enableshadow --enablemd5
selinux --disabled
timezone America/New_York
bootloader --location=mbr --driveorder=xvda
reboot
zerombr yes
clearpart --all
part /boot --fstype ext3 --size=100
part swap --size=256
part pv.00 --size=1 --grow --asprimary
volgroup rootvg pv.00
logvol / --name root_lv --vgname=rootvg --size=400
logvol swap --name swap_lv --vgname=rootvg --size=256
logvol /usr --name usr_lv --vgname=rootvg --size=2000
logvol /tmp --name tmp_lv --vgname=rootvg --size=200
logvol /var --name var_lv --vgname=rootvg --size=200
logvol /home --name home_lv --vgname=rootvg --size=400
%packages
@admin-tools
@base
@core
@dns-server
@editors
@text-internet
@web-server
comps-extras
cracklib-dicts
gnome-mime-data
rmt
tzdata
%post
--
* The Digital Hermit http://www.digitalhermit.com
* Unix and Linux Solutions kwan(a)digitalhermit.com
18 years, 3 months
FC5 installation notes
by Kwan Lowe
Just posting my installation notes here in case someone finds them useful:
Installation media setup:
Setup FTP
yum -y install vsftpd
service vsftpd start
Loopback mount the installation DVD to /var/ftp/pub
mount -o loop /export/C-5-i386-DVD.iso /var/ftp/pub
Create a VBD for the installation
lvcreate -L 3G -n xm-fc5-001_lv rootvg
Run the installation
xenguest-install.py -n xm-fc5-001 \
-f /dev/rootvg/xm-fc5-001_lv \
-r 256 -l ftp://192.168.0.10/pub/
Caveats:
Installer doesn't check if enough memory is available. Drop to run level 3 and/or
mem-set other existing dom0s to lower memory (or shut them off).
Installation path is to the root of the DVD, not the Yum repository package path.
Copying an LVM (easy way to duplicate VBD)
1) Start with an installed DHCP enabled domU.
2) lvcreate -L 2G -n copyof_xm rootvg
3) cd /dev/mapper; dd if=rootvg-xm of=rootvg-copyof--xm bs=1M
4) cp /etc/xen/xm /etx/xen/copyof_xm
5) edit /etc/xen/copyof_xm; change the MAC, name
6) xm create copyof_xm
--
* The Digital Hermit http://www.digitalhermit.com
* Unix and Linux Solutions kwan(a)digitalhermit.com
18 years, 3 months
xen on fc5
by tournesol
Hi again
I hava another guestions again.
Q1: The xenguest-install.py was modified by FC5 team ? or xen team ?
Can i run it on FC4 (or FC2,3) with xen-3.0.1-4 later?
Q2: What is the difference between DomU kernel and Dom0 kernel ?
Which should i run ?
Q3: Where is the anaconda comes ?
when i install the guest-os via VNC anaconda , the anaconda comes from
host-os ? or from xen?
thanks for your help.
tournesol
18 years, 3 months
Page space questions
by Kwan Lowe
Hello All:
I've been successfully running several domains a test box under various workloads
for about 48 hours now. Domains include an httpd server, mysql database, DNS, and
development domain.
My question is about memory sizing and page space. On a physical box one would
allocate an appropriate page space on the local disks. In the DomainU's, however,
would there be any advantage to maybe disabling page space entirely and allocating a
larger base memory from the host? Will the Domain0 be able to allocate more
physical memory than is available by using page space?
E.g.
The physical box (1G RAM) currently has four 128-192M domains with 128-256M page
space defined in each DomainU. DomainU's are backed by LVM devices on a single 200G
physical disk. I would like to have four domains of 256M apiece. This is greater
than the available physical RAM, but at any one point the total memory in use by the
DomainUs is much less. I'd probably have to tweak the individual overcommit settings
in each DomainU.
Anyhoo, the rationale is that the Domain0 may be better able to deal with disk I/O
rather than each DomainU hitting the disk...
If this is fatally flawed reasoning, another thought would be to have a monitor
process on each DomainU report back to the Domain0. It would send memory commit
information which the Domain0 could use to mem-set the DomainU's lower or higher.
I'm thinking of a dynamic memory allocation scheme that could better utilize
available physical memory.
Thoughts? Suggestions? Derision?
(BTW, xen just rocks. It compares very favorably with the VMWare and AIX
micropartitions I use at my day job :D ).
--
* The Digital Hermit http://www.digitalhermit.com
* Unix and Linux Solutions kwan(a)digitalhermit.com
18 years, 3 months