virt March 2008

virt@lists.centos.org

24 participants
24 discussions

Re: [CentOS] Running CentOS 4.6 domUs on CentOS 5.1 dom0 and domUs crash
by Tim Verhoeven 22 Mar '08

22 Mar '08

On Mon, Mar 17, 2008 at 8:05 PM, Nathan Grennan <centos(a)cygnusx-1.org> wrote: > I have a CentOS 5.1 dom0 running on a machine with two cores and 4gb > of memory. It runs three 4.6 domUs and one 5.1 domU. Just out of the > blue the other night all the 4.6 domUs crashed, but the dom0 stayed up. > I ran xm console (domU-name) and got the Oops information below. I > didn't get anything from the 5.1 domU. I think it had been rebooted a > few days before and hadn't come up because of a initrd issue. > > First 4.6 domU: > > kernel BUG at arch/i386/mm/pgtable-xen.c:306! ...snip... Hi, I'm moving this thread to the CentOS-virt mailinglist (please remove the centos mailinglist to everyone replying to this mail) since it is more suited for these kind of questions. Question, If I read those crashes correctly your 4.6 domU is a 32 bit one ? What platform is your dom0 32 or 64 bit? Running a 32 bit domU on a 64 bit dom0 is currently not really stable. This is suppose to improve with 5.2 Regards, Tim -- Tim Verhoeven - tim.verhoeven.be(a)gmail.com - 0479 / 88 11 83 Hoping the problem magically goes away by ignoring it is the "microsoft approach to programming" and should never be allowed. (Linus Torvalds)

8 9

Xen Windows Clients - BSOD with Application Firewall Installs
by centos-virt＠mymail.isbest.biz 21 Mar '08

21 Mar '08

I have set up both Windows XP and Windows 2000 Server clients on Xen under Centos 5.1 and am unable to install an application firewall. I have tried the latest: 1) Agnitum Outpost Commercial Version 2) Comodo Firewall (ver 2 on Windows 2000) 3) Sunbelt Commercial Version With all of these when the firewall starts up it results in a BSOD. With Windows XP and Agnitum I can go back to a previous system restore point, with Windows 2000 and other firewalls on XP the only option is a backup of the Xen client as even a save mode command prompt startup results in a BSOD. I have tried on the Windows XP install using various HALS and switches for PAE ACPI and multi/single cpu support and have not had any success. Windows 2000 BSOD results from "an attempt was made to write to read only memory" NDIS.SYS Windows XP BSOD message is related to: Agnitum - sandbox.sys Others: "an attempt was made to write to read only memory" NDIS.SYS Has anyone any idea of how to set up Xen Windows clients to allow an application firewall to be installed successfully? Thanks Jesse

1 0

virt-install in CentOS 5
by Kanwar Ranbir Sandhu 21 Mar '08

21 Mar '08

Hi All, [I originally posted this message in the main CentOS users' list, but later found this list. So, I'm reposting it here since this is where it belongs.] I'm trying to install a Fedora 8 domU on a CentOS 5 dom0, but the following command fails: [root@xenmaster ~]# virt-install -p -n krs -r 512 -m 00:16:3E:66:94:5F -b xenbr0 --vnc --vncport=5910 -l http://10.19.75.200/mirror/fedora/linux/releases/8 -f /dev/vg00/krs -d Starting install... Sat, 08 Mar 2008 12:27:56 DEBUG Fetching URI http://10.19.75.200/mirror/fedora/linux/releases/8//Fedora Sat, 08 Mar 2008 12:27:56 DEBUG Saved file to /var/lib/xen/virtinst-Fedora.VhrSOC Retrieving Fedora... 100% |=========================| 965 B 00:00 Sat, 08 Mar 2008 12:27:56 DEBUG Detected a Fedora distro Sat, 08 Mar 2008 12:27:56 DEBUG Fetching URI http://10.19.75.200/mirror/fedora/linux/releases/8/images/xen/vmlinuz ERROR: Invalid URL location given: [Errno 14] HTTP Error 404: Not Found I have no idea why virt-install is looking for the xen kernel in http://10.19.75.200/mirror/fedora/linux/releases/8/images/xen/vmlinuz : it's actually here: http://10.19.75.200/mirror/fedora/linux/releases/8/Fedora/i386/os/images/xe… I've used the above syntax to install CentOS 5 domUs on the same host without any problems. What gives? Regards, Ranbir -- Kanwar Ranbir Sandhu Linux 2.6.22.14-72.fc6 i686 GNU/Linux 20:51:42 up 4 days, 9:58, 2 users, load average: 1.23, 0.63, 0.48

6 15

Re: [CentOS-virt] RE: [CentOS] Running CentOS 4.6 domUs on CentOS5.1 dom0 and domUscrash
by Ross S. W. Walker 21 Mar '08

21 Mar '08

Sorry for the top post, blackberry. Well I like to be able to install my machines in the xenstore with 'xm new' and then manage them from there, but virt-install and virsh are useful too, so I just use CentOS base then replace their xen with the one from xen.org keeping libvirt and now I have the best of both worlds. The commercial one isn't without gripes too. Like I wish they had developed a fully web based interface for creating and managing virtual machines like the one Microsoft has in their virtual server instead of the standalone java client. I never really like standalone java clients as it makes it one more thing you have to reach for when you need to get something done, when there is almost always a web browser available. -Ross ----- Original Message ----- From: centos-virt-bounces(a)centos.org <centos-virt-bounces(a)centos.org> To: Discussion about the virtualization on CentOS <centos-virt(a)centos.org> Sent: Thu Mar 20 17:34:08 2008 Subject: Re: [CentOS-virt] RE: [CentOS] Running CentOS 4.6 domUs on CentOS5.1 dom0 and domUscrash Ross S. W. Walker wrote: > glues the Xen 3.0.3 user land tools onto a highly patched 3.1 > hypervisor forcing the user to try and use their swiss army > knife libvirt to manage the mess. Ugh! I just want Xen and > to manage it through Xen! I quite like the virsh and virt-install stuff included in the distro, I think xen tools themselves are quite pedestrian at the moment. Perhaps their commercial offerings are better - for me, the libvirt and virsh make Xen a lot more usable -- Karanbir Singh : http://www.karan.org/ : 2522219@icq _______________________________________________ CentOS-virt mailing list CentOS-virt(a)centos.org http://lists.centos.org/mailman/listinfo/centos-virt ______________________________________________________________________ This e-mail, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this e-mail, you are hereby notified that any dissemination, distribution or copying of this e-mail, and any attachments thereto, is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender and permanently delete the original and any copy or printout thereof.

1 0

Clock fast - work around
by chanms 20 Mar '08

20 Mar '08

Well, to the clock being fast work around, I kind of found a way to work around it for now... 1) use a distribution that supports VMI (VMware's paravirtualization) 2) enable VMI paravirtualization support in VMware Workstation 6 VM's profile 3) use vmi-timer as the clocksource (as opposed to pit, acpi_pm, etc.) 4) CPU must be constantly at maximum frequency For 1), I use Ubuntu 7.10 For 2), the VM must be VMWare 6 profile type, then in Settings -> Options -> Advanced, check on "Enable VMware paravirtual kernel support"). I was unable to use VMI before because my entire Vista (x64 Ultimate) machine will BSOD when Ubuntu was booted with VMI turned on; only in the recent Workstation 6.0.3 patch that this bug has been fixed. For 3), Ubuntu default to vmi-timer if 1) and 2) are working properly; can be verified by /sys/devices/system/clocksource/clocksource0/available_clocksource and current_clocksource For 4), I didn't want to constantly lock the CPU of my laptop at highest speed at any time, as I don't need the CPU at full blast when I'm browsing website, or using MSOffice. So what I've done is by using distributed.net - this program uses idle CPU power to do calculations. This program keeps my CPU busy at all times but easily switched out when other programs need the CPU. With this, the Ubuntu's VM clock runs at a fairly accurate rate. The side effect is my CPU is constantly busy and the laptop cooling fan kept going and going (and no more putting my PC on my lap as it will be fairly hot). I also have a desktop with AMD Athlon X2 running Windows 2003 Server; with the above configuration (minus #4, as W2K3 Server keeps CPU clock rate constant) the Ubuntu VM's clock also was within 5 seconds compared to the host PC's clock over the period of a day. I hope that VMware will be able to devise a better way to keep VM clocks accurate in the future in their products.

1 0

help getting open-vm-tools installed
by Fletch Bowling 19 Mar '08

19 Mar '08

Hello, Newbie to CentOS and attempting to run it under a windows xp host. I see there is a open version of vmware tools available but getting various dependency errors when attempting to install. the rpms. Will these tools install for ver CentOS 5 or are they for an earlier version? P.S. How closely does CentOS match Redhat binaries? VMware tools (from workstation 6) install just fine but could not be configured due to various errors. Thanks, Fletch

2 7

Follow-up on Fast clock under VMWare
by chanms 16 Mar '08

16 Mar '08

Hi everyone: Just a follow up on the fast clock issue: "Can you double-check that 'cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq' outputs the same number as your 'cpuKHZ' setting in your config.ini file?" There is no cpufreq directory under /sys/devices/system/cpu/cpu0; all I see is "cache" (dir), "crash_notes" (a file) and "topology" (dir). However under /sys/devices/system/clocksource/clocksource0, available_clocksource reads acpi_pm jiffies tsc pit, current_clocksource reads pit. (I put in the line clocksource=pit) in the grub boot up line. Divider=10 in the bootup line causes the boot up process to hang and a reset has to be performed; removing it and the CentOS boots properly. I'm currently using 2.6.18-53.1.13.el5, with boot up options are "noapic nolapic quiet clocksource=pit" (took out divider=10). CentOS clock is going at around 2 seconds per every real second.

3 2

Fwd: 2 problems running CentOS 5.1 + kernel-vm in VMware Fusion
by Leonardo Vilela Pinheiro 08 Mar '08

08 Mar '08

Hi, Inside VMware Fusion 1.1.1, I have a virtual machine running CentOS 5.1 x86_64 with: - kernel-vm-2.6.18-53.1.13.el5.x86_64.rpm from http://people.centos.org/~tru/kernel-vm/5/RPMS/x86_64/ - kmod-open-vm-tools-0-0.20080123.2.6.18_53.1.13.el5vm.x86_64.rpm from http://people.centos.org/~hughesjr/open-vm-tools/5/x86_64/ First problem: Sometimes (most times) I get this error during boot (first error): "The VMware Tools poweron script did not run successfully in this virtual machine - If you have configured a custom poweron script in this virtual machine, please inspect it for errors; otherwise, please submit a support request." A screenshot of this error can be seen in: http://www.scribd.com/doc/2228546/VMware-Fusion-error-running-CentOS-51-ker… This problem is very strange, it doesn't happen on every boot, and usually when I boot another kernel (like change kernel-vm to "kernel-default" or back) it disappears and only appears again after some reboots using the kernel-vm. I don't know a safe path to reproduce this bug, and I also don't know for sure if it only happens with kernel-vm. Second problem: When I try to suspend the virtual machine (clicking on "Suspend Guest"), I get the error that can be seen in: http://www.scribd.com/doc/2232883/VMware-Fusion-running-CentOS-51-kernelvm-… When I try to suspend again, it gives no error and suspends, but when it resumes it gives the error that can be seen in: http://www.scribd.com/doc/2233720/CentOS-51-in-VMware-Fusion-resuming-from-… The second problem also happens when using: - kernel-2.6.18-53.1.14.el5 (default kernel) - kmod-open-vm-tools-0-0.20080123.2.6.18_53.1.13.el5vm.x86_64.rpm The second problem does *not* happen when using: - kernel-2.6.18-53.1.14.el5 (default kernel) or kernel-vm-2.6.18-53.1.13.el5.x86_64.rpm - default vmware-tools (which comes with VMware Fusion 1.1.1). So it might be a problem in kmod-open-vm-tools. If I may help debugging something, please tell me what to show you. -- Leonardo

1 1

Multi domU Xen Box Network Transparent Sound
by Ross S. W. Walker 08 Mar '08

08 Mar '08

I have a new development box at home that I am working on. Currently it has CentOS 5.1 + Xen 3.2 running several domUs. I am using XDMCP on the domUs and an X Chooser on dom0 for connecting to them and it works well, but now I want to be able to get sound from my domUs and I have been struggling. I can get some from network transparent sound by using ESD, but I have no way of controlling the mixer (volume) settings from within the domU which kind of sucks. I have been looking at the possibility of setting up a PulseAudio server in dom0 and installing the ALSA PulseAudio plugin on each domU set to use the dom0 server as from the web site it seems to be able to have remote mixer controls and can be handled as if it were a regular ALSA sound card, which should allow a majority of the apps to work as if there were a local sound card. My questions are: 1) Has anybody on the list used PulseAudio in this type of setup (not necessary Xen interface, but ALSA plugin to remote PulseAudio server)? 2) Does anybody on the list know a compatible repo that carries pre-compiled PulseAudio packages? Thanks in advance, Ross S. W. Walker Information Systems Manager Medallion Financial, Corp. 437 Madison Avenue 38th Floor New York, NY 10022 Tel: (212) 328-2165 Fax: (212) 328-2125 WWW: http://www.medallion.com <http://www.medallion.com/> ______________________________________________________________________ This e-mail, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this e-mail, you are hereby notified that any dissemination, distribution or copying of this e-mail, and any attachments thereto, is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender and permanently delete the original and any copy or printout thereof.

1 0

Question on best way of configuring this
by Stephen Harris 03 Mar '08

03 Mar '08

I have a Windows XP machine that I don't use very often, but I leave it powered up because once in a blue moon I need to remotely access it (eg from work), and to keep the Antivirus software up to date. When I do use it at the console I'm mostly doing basic video editting or transcoding (eg from transferred TiVo content). I sometimes play older games on it (eg Carmaggedon or Worms 2; I'm not into Warcrack, or City Of Heros or anything that needs a super-duper gaming card). I also have an aging Linux machine. This is my main workstation, although I primarily access it as a server (ssh into it, screen). I do occaisionally use the console as an X display, eg when scanning stuff or previewing my nroff documents before sending them to the printer. This machine has all my email, runs leafnode etc etc. What I'd like to do is run them both as virtual instances on a new machine (quad core duo) I've ordered. What I would really like to have is something like the dom0 being on tty0, map the Linux domU consoles to tty1->5, allow the Linux domU X server to run fullscreen on tty6, and have the windows GUI appear fullscreen on tty7. I really don't want to remotely access the windows GUI (it makes video editting harder, and game playing almost impossible), and would want to avoid displaying it inside another X session. Ideally it'd feel like a "real" machine. Does this sound like a feasible option? -- rgds Stephen

2 1

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

virt March 2008