xen May 2006

xen@lists.fedoraproject.org

69 participants
92 discussions

Easiest way to compile custom kernel, for fedora +openswan+xen+patchomatic
by Tom Bishop 30 May '06

30 May '06

I want to compile a custom xenU kernel for a firewall distribution (endianfw, version of ipcop), I can get it to load up but the kernel is missing some modules that aren't resident. So I want to make a custom version for this domU, I know what I need kernel+openswan+xen+patch-o-matic but I not sure the best way to go about this, do I patch a fedora source or start with a vanilla kernel? I know what I want but am unsure what is the best way? I have read that this has been done under cool configurations but they didn't go into much detail. Any guidance would be appreciated. Thanks.

3 4

DomU problems with LVM
by Gawain Lynch 30 May '06

30 May '06

After a lot of testing and fun I have freed a box for production Xen installs. When I try to install Xen onto an LV things go quite wrong. If I create and mount an LV and install Xen into a file backing then all works fine. The process I am using goes like this... Create an LV: # lvcreate -L 2G vgxen -n myxen Install FC5: # xenguest-install.py --name=myxen \ --file=/dev/vgxen/myxen \ --ram=256 \ --location=http://install/5/i386/ I preform a minimal install and it finishes without errors, the only thing I customise is the name of the volume group (VolGroup00 => vg00) and the logical volumes (LogVol00 => lv00 etc), I'm not a fan of the StudlyCaps default :-) The I start the instance # xm create -c myxen Only to be confronted with this: Setting up Logical Volume Management: 2 logical volume(s) in volume group "vg00" now active [ OK ] Checking filesystems Checking all file systems. [/sbin/fsck.ext3 (1) -- /] fsck.ext3 -a /dev/vg00/lv00 /dev/vg00/lv00: clean, 19347/418080 files, 126997/417792 blocks [/sbin/fsck.ext3 (1) -- /boot] fsck.ext3 -a /dev/xvda1 Error reading block 32 (Attempt to read block from filesystem resulted in short read). /boot: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY. (i.e., without -a or -p options) Trying to fsck the /boot filesystem gives the following errors: Error reading block 32 (Attempt to read block from filesystem resulted in short read). Ignore error<y>? Does anyone have any ideas as to what is wrong or what I am doing wrong? Take care, Gawain

1 0

FC5/Xen LVM resize question
by Heinz Deinhart 30 May '06

30 May '06

Hi, im playing around with Xen and FC5. There is one thing i dont get working. When i resize a logical volume in dom0 with lvextend, and this volume is used and mounted in a domU, it seems that the domU does not notice the new size of the partition until rebooted. Should this work somehow without rebooting the domU, or is this the intended behavior? I think, i could create a volume goup inside of a domU and add space to the domU by adding new block devices and adding their space to the volume group. But is this really a clever way of doing this, or are there any easier ways? Thanks a lot, Heinz -- Heinz Deinhart emp(a)nix.at

2 2

Questions about virtualization and Xen
by Filipe Miranda 29 May '06

29 May '06

Hello everybody, I would like to clear a few questions that I have about virtualization, please correct me if I'm wrong... 1) Xen is a virtual machine monitor that uses the hardware virtualization approach knows as para-virtualization. 2) Para-virtualization is a technology that provides the virtual machine monitor access directly to the machine's hardware without having an underlying operation system to manage that, so less layer of abstraction are between the virtual machines OS guest and the true hardware. 3) Xen 3.0 makes possible to guests virtual machines to access the machine's hardware either through Xens APIs or through hardware enabled virtualization like Intel VT-x or AMD Pacifica technologies, so you can have either a modified guest OS like Linux or an unmodified OS guest like Windows2003 running on top of Xen 3.0. Are those thoughts okay? -- Att. --- Filipe

2 1

RE: IPF DomVTi BIOS image rpmfor FC6/IPF/Xen is available.
by Yu, Fenghua 27 May '06

27 May '06

The last email is too big. It's in the approval process by Redhat. In case it won't get to this mailing list. I'll send it out in other way. Thanks. -Fenghua _____________________________________________ From: Yu, Fenghua Sent: Friday, May 26, 2006 6:15 PM To: 'fedora-xen(a)redhat.com' Subject: IPF DomVTi BIOS image rpmfor FC6/IPF/Xen is available. Please check the attached source rpm and binary rpm for IPF VTi BIOS image rpmfor FC6/IPF/Xen. This is used to install BIOS image for DomVTi. Thanks. -Fenghua << File: xen-ia64-flash-1.0.0-1.ia64.rpm >> << OLE Object: Outlook File Attachment >>

1 0

RE: [Fedora-xen] FW: [IA64]We have successfully booteddom0/domU/domVTI (xen/ia64) on FC5
by Zhang, Xiantao 26 May '06

26 May '06

> Hi > > > > - Is it possible to use same configuration file for xen0/xenU > > > just like current xen-ia64-unstable.hg? Or at least give > > > a configuration option to user in spec. > > > > Right now the build follows the pattern of the other architectures in > > the spec file. If Juan is interested in pursuing a -xen kernel > > instead of -xen0/-xenU, I'd be happy to go that route. > > I am studing doing that, but domU already gives enough errors/warnings > (while booting, things like rtc not available and somesuch). It should exist in native xen-ia64-unstable. > But yes, I will be way, way happier with a single kernel, working on > that :p Maybe reference the xen-ia64-unstable way is easier. Thanks Xiantao

1 0

FW: [IA64]We have successfully booted dom0/domU/domVTI (xen/ia64) on FC5
by Zhang, Xiantao 26 May '06

26 May '06

Due to picture size exceeding limitation of mailing list, it's waiting for approval now. So, send out notification first for your reference. :-) Thanks -Xiantao -----Original Message----- From: Zhang, Xiantao Sent: 2006年5月26日 16:24 To: fedora-xen(a)redhat.com Subject: [IA64]We have successfully booted dom0/domU/domVTI (xen/ia64) on FC5 Hi, all, After some hard debug and work around, now we've got xen/ia64 dom0/domU/domVTI both up on FC5-IA64. Attached is the snapshot. We still use python2.4.2 contained in FC5 for test. We need following changes to make things happen: - xen-tools is not built and integrated into the RPM package. Typically we know there're 3 major directories on xen-ia64-unstable.hg: xen,linux-sparse and tools. All three need to be kept consistence with each other. However current fedora-kernel-ia64 only incorporates former two and tools is totally missed. We tried to build the tools individually from xen-ia64-unstable under same Rev (10150), and then install to target box manually. - The configuration files for dom0/domU seems a big difference compared to xen-ia64-unstable.hg. (Why?) Currently the ext3 is compiled as module, however initrd doesn't work for domU even when we add "ramdisk=" in config file. So we change ext3 to be kernel built-in, and then domU can boot up immediately. * Xen-ia64-unstable.hg doesn't have IDE devices configured for domU, however fedora-kernel-ia64 does, which wastes much time for probe. If the configuration file can be kept consistence with xen-ia64-unstable.hg, above issues are gone. After above 2 changes, both domU and domVTI can be boot successfully. Following are other tricky issues we've observed: - brctl package is not integrated into FC5 and so vnif is not available now. People need to install brctl manually and we'll try network later. Is it possible to add dependency check upon brctl in spec file? - When installing xen0 rpm package, it depends on xen rpm package. However we didn't find the xen rpm package, and had to install with "nodeps". Actually xen image has been included in xen0 rpm package. - Is it possible to use same configuration file for xen0/xenU just like current xen-ia64-unstable.hg? Or at least give a configuration option to user in spec. - Even everything works now, we kept seeing some boring call trace from BUG in mm/slab.c:2818, especially when vncserver is started. Actually even dom0 itself will encounter such warning infrequently before xend start. Cheers! -Xiantao and Kevin

3 3

Anaconda fails when initializing install
by Ronny 26 May '06

26 May '06

Hi folks am trying to install a Core 5 guest on a Core 5 dom0 box.Install location is http//*** on the same box and that's where I copied the Fedora tree off the iso cd.All goes well up to the package installation stage and reports bugs in anaconda which I don't think is the cause.Have 512MB of RAM and saved 128MB for the dom0 + 4gigs to the guest box :'( Any hints please are welcome? Thx Ronny -- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- If I have seen further it is by standing on the shoulders of giants. --Isaac Newton -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

2 2

RE: [Fedora-xen] FW: [IA64]We have successfully booteddom0/domU/domVTI (xen/ia64) on FC5
by Zhang, Xiantao 26 May '06

26 May '06

Hi Juan, I am not sure your release version is same to mine. I didn't find the rpm package in my box although I chose the full installation. Anyway , I have installed sysfs-util and bridge rpm packaged successfully, seems network is OK for me. Thanks for your care. -Xiantao > -----Original Message----- > From: Juan Quintela [mailto:quintela@redhat.com] > Sent: 2006年5月26日 19:16 > To: Zhang, Xiantao > Cc: fedora-xen(a)redhat.com > Subject: Re: [Fedora-xen] FW: [IA64]We have successfully > booteddom0/domU/domVTI (xen/ia64) on FC5 > > Hi > > > Hi, all, > > After some hard debug and work around, now we've got xen/ia64 > dom0/domU/domVTI both up on FC5-IA64. Attached is the snapshot. > > Congrats > > > Following are other tricky issues we've observed: > > - brctl package is not integrated into FC5 and so vnif is not available > now. People need to install brctl manually and we'll try network later. Is it > possible to add dependency check upon brctl in spec file? > > Ein? > [quintela@trasno ~]$ rpm -qf /usr/sbin/brctl > bridge-utils-1.0.6-1.2 > [quintela@trasno ~]$ cat /etc/fedora-release > Fedora Core release 5 (Bordeaux) > [quintela@trasno ~]$ > > Or do you mean that it is not integrated in fedora? > > Later, JUan.

1 0

[IA64]We have successfully booted dom0/domU/domVTI (xen/ia64) on FC5
by Zhang, Xiantao 26 May '06

26 May '06

Hi, all, After some hard debug and work around, now we've got xen/ia64 dom0/domU/domVTI both up on FC5-IA64. Attached is the snapshot. We still use python2.4.2 contained in FC5 for test. We need following changes to make things happen: - xen-tools is not built and integrated into the RPM package. Typically we know there're 3 major directories on xen-ia64-unstable.hg: xen,linux-sparse and tools. All three need to be kept consistence with each other. However current fedora-kernel-ia64 only incorporates former two and tools is totally missed. We tried to build the tools individually from xen-ia64-unstable under same Rev (10150), and then install to target box manually. - The configuration files for dom0/domU seems a big difference compared to xen-ia64-unstable.hg. (Why?) Currently the ext3 is compiled as module, however initrd doesn't work for domU even when we add "ramdisk=" in config file. So we change ext3 to be kernel built-in, and then domU can boot up immediately. * Xen-ia64-unstable.hg doesn't have IDE devices configured for domU, however fedora-kernel-ia64 does, which wastes much time for probe. If the configuration file can be kept consistence with xen-ia64-unstable.hg, above issues are gone. After above 2 changes, both domU and domVTI can be boot successfully. Following are other tricky issues we've observed: - brctl package is not integrated into FC5 and so vnif is not available now. People need to install brctl manually and we'll try network later. Is it possible to add dependency check upon brctl in spec file? - When installing xen0 rpm package, it depends on xen rpm package. However we didn't find the xen rpm package, and had to install with "nodeps". Actually xen image has been included in xen0 rpm package. - Is it possible to use same configuration file for xen0/xenU just like current xen-ia64-unstable.hg? Or at least give a configuration option to user in spec. - Even everything works now, we kept seeing some boring call trace from BUG in mm/slab.c:2818, especially when vncserver is started. Actually even dom0 itself will encounter such warning infrequently before xend start. Cheers! -Xiantao and Kevin

1 0

← Newer
1
2
3
4
5
6
7
8
9
10
Older →

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

xen May 2006