virt

virt@lists.centos.org

1 participants
1609 discussions

PCI passthrough problems
by Enderson Maia 23 Sep '08

23 Sep '08

Hello, I'm having a problem, I've seen other people with the same problem, but with USB. Some people solved this with the pciback.permissive option, but in my case it's ignored. I'm using CentOS 5.2 - 2.6.18-92.1.10.el5.centos.plusxen The error return from the console starting the domU is below. I've tried swiotlb=force, pciback.permissive like I said. My configs are: /etc/modprobe.conf ... options pciback hide=(03:01.0) # cat /etc/xen//firewall name = "firewall" uuid = "cd5f6ad8-0682-73fd-3268-fb4a7b85c0ec" maxmem = 256 memory = 256 vcpus = 1 bootloader = "/usr/bin/pygrub" boot = "swiotlb=force" on_poweroff = "destroy" on_reboot = "restart" on_crash = "restart" vfb = [ "type=vnc,vncunused=1" ] disk = [ "phy:/dev/VolGroup00/firewall.disk,xvda,w" ] vif = [ "mac=00:16:3e:0e:53:5a,bridge=virbr0" ] pci = [ '0000:03:01.0' ] # rpm -qa | grep xen xen-libs-3.0.3-64.el5_2.1 xen-3.0.3-64.el5_2.1 kernel-xen-2.6.18-92.el5 kernel-xen-2.6.18-92.1.10.el5.centos.plus --- Iniciando interface eth0: [ OK ] Iniciando interface eth1: Determinando informação IP para o eth1...Fatal DMA error! Please use 'swiotlb=force' ------------[ cut here ]------------ kernel BUG at arch/i386/kernel/pci-dma-xen.c:401! invalid opcode: 0000 [#1] SMP last sysfs file: /class/net/eth1/address Modules linked in: xennet ip_conntrack_netbios_ns ipt_REJECT xt_state ip_conntrack nfnetlink iptable_filter ip_tables ip6t_REJECT xt_tcpudp ip6table_filter ip6_tables x_tables ipv6 xfrm_nalgo crypto_api dm_mirror dm_multipath dm_mod parport_pc lp parport 3c59x pcspkr mii xenblk ext3 jbd uhci_hcd ohci_hcd ehci_hcd CPU: 0 EIP: 0061:[<c040a39c>] Not tainted VLI EFLAGS: 00010292 (2.6.18-92.el5xen #1) EIP is at dma_map_single+0x12d/0x147 eax: 0000002f ebx: 00000000 ecx: 00000001 edx: f5416000 esi: 38f59e02 edi: 00000001 ebp: 0000004e esp: c071bdbc ds: 007b es: 007b ss: 0069 Process swapper (pid: 0, ti=c071b000 task=c066e2c0 task.ti=c06e6000) Stack: c0621abd c07fc448 008a8e02 00000000 cf914400 cf5db200 c07fc448 00000140 d1232e6e 00000001 c08a8e02 cf914000 cf88c200 00000002 cf5db200 0000004e d126c000 cf914000 0000004e c08a8e02 cf88c140 cf88c140 cf88c200 c07c4c88 Call Trace: [<d1232e6e>] boomerang_start_xmit+0x141/0x352 [3c59x] [<c05aec59>] dev_hard_start_xmit+0x198/0x1ee [<c05bcaf6>] __qdisc_run+0xe4/0x19a [<c05b0544>] dev_queue_xmit+0x1e0/0x2e8 [<c05b4b24>] neigh_resolve_output+0x1b2/0x1dc [<d136eb2e>] ip6_output2+0x256/0x2a7 [ipv6] [<d136f656>] ip6_output+0x9c7/0x9f3 [ipv6] [<d137c8c0>] dst_output+0x0/0x7 [ipv6] [<c05c3d12>] nf_hook_slow+0x3a/0x90 [<d137c8c0>] dst_output+0x0/0x7 [ipv6] [<d137e4d6>] ndisc_send_ns+0x3a8/0x47a [ipv6] [<c06090d8>] _spin_lock_irqsave+0x8/0x28 [<c042914a>] lock_timer_base+0x15/0x2f [<d1375ddb>] addrconf_dad_timer+0xc3/0xdf [ipv6] [<c0434244>] hrtimer_run_queues+0x14b/0x163 [<d1375d18>] addrconf_dad_timer+0x0/0xdf [ipv6] [<c0429600>] run_timer_softirq+0x101/0x15c [<c042613e>] __do_softirq+0x5e/0xc3 [<c0406edf>] do_softirq+0x56/0xaf [<c0406e80>] do_IRQ+0xa5/0xae [<c0549913>] evtchn_do_upcall+0x64/0x9b [<c04055d9>] hypervisor_callback+0x3d/0x48 [<c0408632>] raw_safe_halt+0x8c/0xaf [<c040321a>] xen_idle+0x22/0x2e [<c0403339>] cpu_idle+0x91/0xab [<c06eb9f5>] start_kernel+0x37a/0x381 ======================= Code: c0 74 07 8b 08 8b 58 04 eb 05 83 c9 ff 31 db 89 da 89 c8 f7 d2 f7 d0 21 fa 21 f0 89 d1 09 c1 74 13 68 bd 1a 62 c0 e8 de 7b 01 00 <0f> 0b 91 01 9e 1a 62 c0 5d f0 83 04 24 00 83 c4 0c 89 f0 89 fa EIP: [<c040a39c>] dma_map_single+0x12d/0x147 SS:ESP 0069:c071bdbc <0>Kernel panic - not syncing: Fatal exception in interrupt --- If you need more information, just tell -- Enderson Maia

1 0

Status of kernel divider option in 2.6.9-78.0.1
by Jeff 20 Sep '08

20 Sep '08

Under 4.6, we recompiled the kernel with HZ=100 for improved time-keeping in VMware guests. I've read about the backporting of the divider patch into RHEL/CentOS 4.7, but it sounds like it also comes with some bugs. I have been unable to determine the current status of the divider option in the latest 4.7 kernel update. I have experimented with "divider=10" and it works with no major problems, except that the guest clock drifts ahead very slightly. So far, it looks like several seconds of drift per day. Not a show-stopper, but still a concern. With HZ=100, our time sync was spot on. In both cases we are using "clock=pit nosmp noapic" and tools.syncTime="TRUE". If it matters, our host machines are CentOS 4.6 x86_64. Can anyone comment on the current status of bugs with the divider option? Is anyone getting accurate time sync with divider=10 with 2.6.9-78.0.1? Is there anything else to watch out for? Is there any fix for the slight time drift with divider=10? I am primarily concerned with i686. Currently, the SRPM for -78.0.1 is missing from the mirrors, so if I want to update, I have to rely on the divider option for my VMware guests. It also appears that Tru has not yet added any -78 RPMs to his kernel-vm project. -- Jeff

2 4

vmware on centos cluster
by Mattias Hemmingsson 11 Sep '08

11 Sep '08

Hello I have one question about cluster and vmware. I have about 10 computers they are all old once from 1g and 256 ram. And now im thinking of putting them all toghter in one cluster. This cluster should be an high performing so all computers share they performance. Now on this cluster i whould like to install one vmware server and make that vmware server get the power from all these computers. And install a os in that vmware server. Is this the way passible to do ? // matte

2 1

Xen kickstart and virt-manager
by Brett Worth 11 Sep '08

11 Sep '08

Hi again. I'm pleased to say that my experimentation with ocfs2 has been very successful and I'm considering doing some real work with it. My current question is related to virt-manager and the use of kickstart to install paravirtualised domU's. I'm happily doing fully-virtualised installs with this tool but due to poor network performance I thought it time to do some para-virtualised testing. In the dialog where I'm asked to supply the URL for the install repository and the kickstart location I'm unable to use the "file://..." construct so I'm left with a selection that only includes locations on the network. Now my network interface information for the domU is in the kickstart file so when I move forward in the tool there's a message saying that its trying to determine a network address, presumably via DHCP (I am unable to put a dhcp server on this network). From the command line I can use wget against the kickstart file at the URL I'm providing so that's OK. I'm obviously overlooking something but its currently got me puzzled. So how is this supposed to work? Am I supposed to do the network configuration for the installation somewhere else? Brett

3 4

Disappearing TAP interface
by James Roman 08 Sep '08

08 Sep '08

I have a CentOS 5.2 Xen host, running several guest domains. Each guest domain has a single interface assigned to it. I am using both the network-bridge and vif-bridge scripts. On my only windows guest, the tap interface connected to the bridge keeps getting disabled. If I check "brctl show" the tap0 interface is not present, and there will be several message log entries indicating that the interface tap0 is being disabled. This means that the guest becomes invisible. I can not open a VNC console or connect to it from the network. From the hosts log files, it appears that the OS keeps running, even though it it is unreachable. (this is a little hard to guage, because the times in the eventlog keep jumping around.) The only way I have been able to restore a connection is to destroy the domain and re-create it. I can not find much in the various Xen logs that identifies what triggers the tap interface to be disabled. I have several questions. Is there any way to manually rebuild the tap interface to get access to the host again (either through VNC or network)? Maybe through the virsh command or calling the vif-bridge or qemu-ifup command directly? Is there another way that I could use to maintain a connection to the guest? Does anyone have any idea what causes the tap interface to go down? --qemu-dm log---------------------------------------------------- domid: 1 qemu: the number of cpus is 2 *** image /local/domain/0/backend/vbd/1/768 format raw driver 0x80e7940 *** image /local/domain/0/backend/vbd/1/5632 format raw driver 0x80e7940 *** image /local/domain/0/backend/vbd/1/832 format raw driver 0x80e7940 Watching /local/domain/1/logdirty/next-active xs_read(): vncpasswd get error. /vm/e2a537ca-5ef5-e87d-6505-855529af3ea4/vncpasswd. char device redirected to /dev/pts/1 qemu_map_cache_init nr_buckets = 4000 shared page at pfn 3ffff buffered io page at pfn 3fffd xs_read(/vm/e2a537ca-5ef5-e87d-6505-855529af3ea4/rtc/timeoffset): read error I/O request not ready: 0, ptr: 0, port: 0, data: 0, count: 0, size: 0 I/O request not ready: 0, ptr: 0, port: 0, data: 0, count: 0, size: 0 Triggered log-dirty buffer switch xs_write(/vm/e2a537ca-5ef5-e87d-6505-855529af3ea4/rtc/timeoffset, rtc/timeoffset): write error xs_write(/vm/e2a537ca-5ef5-e87d-6505-855529af3ea4/rtc/timeoffset, rtc/timeoffset): write error --------------------------------------------------------------------------- name = "vm-002_win" uuid = "e2a537ca5ef5e87d6505855529af3ea4" maxmem = 2048 memory = 1024 vcpus = 2 builder = "hvm" kernel = "/usr/lib/xen/boot/hvmloader" boot = "c" pae = 1 acpi = 1 apic = 1 on_poweroff = "destroy" on_reboot = "restart" on_crash = "restart" device_model = "/usr/lib/xen/bin/qemu-dm" sdl = 0 vnc = 1 vncunused = 1 keymap = "en-us" disk = [ "file:/var/lib/xen/images/vm-002_win.img,hda,w", "file:/data/os-install-media/Win2003plusSP1.iso,hdc:cdrom,r", "phy:/dev/VolGroup00/LVssavmWin001Data,hdb,w" ] vif = [ "ip=192.168.2.103,mac=00:16:3e:0d:7f:be,bridge=xenbr0,type=ioemu" ] serial = "pty" -- James D. Roman IT Network Administration Terranet Inc.On contract to: Science Systems and Applications, Inc.

1 0

relocation, migration
by Paras pradhan 02 Sep '08

02 Sep '08

Hi, I am running a cluster having 2 nodes using red hat cluster suite. node 1 has a para virtualized guest(centOS) running under Xen. My question is when node1 is rebooted, guest is automatically relocated to node 2 . Instead of relocation, is migration possible in this case which can result in Zero down time? Thanks in adv Paras.

3 3

xen+drbd Question
by Alexandre Biancalana 27 Aug '08

27 Aug '08

Hi list, I installed two new servers to be our virtualization HA. This servers have CentOS 5.2, Xen 3.0.3, Drbd 8.2.6. The setup is ok, lvm drbd resources sincronizing, Guest VM running without any problem. My only doubt is that I can only use drbd resources inside guest vms when I specify the disk resource in vms config like this: disk = ['phy:/dev/VolGroup00/vappprtem01,hda,w' ] if I specify de disk like: disk = [drbd:vappprtem01,hda,w' ] The Guest VM found the disk but this appear with zero size. As far what I read the second way is the correct to live migration work without any race condiction... This link say that centos has a problem with block-drbd script provided by drbd, is anybody aware of that ? http://lists.xensource.com/archives/html/xen-users/2008-08/msg00176.html Any hints ? Best Regards,

5 13

Re: [CentOS-virt] Shared Storage Options
by Brett Worth 27 Aug '08

27 Aug '08

Jerry Ablan wrote: > Has anyone played with OCFS2? > > http://oss.oracle.com/projects/ocfs2/ Thanks for suggesting this. I've done a fresh install of CentOS 5.2 and put OCFS2 on top. Its really easy to use and install and the the performance looks really good. I'm going to be using OCFS2 for a few more applications from now on. Meanwhile... Once I ran up Xen on these machines I was able to use file based storage allocation to the VMs and have it running with reasonable performance. My original problem is still there however. I've tested the flock functionality of OCFS2 and it works but Xen it seems doesn't do locking even on a shared filesystem. It is possible that Xen is trying to do some other sort of file locking but unfortunately OCFS2 only supports flock. I could write some scripts to replace xm which could check for the existence of a lock file before attempting to start a VM but that seems a bit clunky. Is there a proper way to do this? Regards Brett > -----Original Message----- > From: centos-virt-bounces(a)centos.org [mailto:centos-virt-bounces@centos.org] > On Behalf Of Brett Worth > Sent: Thursday, August 21, 2008 6:22 AM > To: Discussion about the virtualization on CentOS > Subject: [CentOS-virt] Shared Storage Options > > Hello all. > > I would like to canvas some opinions on options for shared storage in a Xen > cluster. So > far I've experimented with using iSCSI and clvm which mixed success. > > The primary concern I have with both of these options is that there seems to > be no obvious > way to ensure exclusive access to the LUN/device to the VM I want to run. > On a couple of > occasions during my playing I've accidentally started the same VM on two > servers and > corrupted the said filesystem. > > I've tried setting up GFS but I don't think I'm smart enough to do that. > The fencing has > me baffled. In fact fencing had me baffled with clvm but I somehow managed > to trick it > into working though probably not how it's supposed to. > > I'd be happy to persist with either of the above methods if I could work out > how to lock > the device. Can anyone offer some advice? > > Regards > Brett Worth

3 3

Installing xen from xen srpm compiled rpms
by Joseph L. Casale 26 Aug '08

26 Aug '08

When I use yum to install xen 3.2.0 and xen-libs it also pulls in then centos kernel-xen? I thought the srpm contained all that was needed besides obvious deps like bridge utils etc? To get this running, I change the grub stanza kernel-xen creates to have the kernel point to xen.gz-3.2 and one of the module lines to point to my custom initrd for pciback loading. That leaves the other module line kernel-xen wrote for vmlinuz-2.6.18-92.1.10.el5xen intact. Why does kernel-xen get pulled in and what's the significance of the vmlinuz line? Thanks! jlc

1 0

VMware Server clock woes (running too fast)
by Michael Ekstrand 25 Aug '08

25 Aug '08

I'm running VMware Server 1.0.6 on a CentOS 5.2 host and am having some clock difficulties. Host OS is x86_64 running on 1.9 GHz AMD Sempron, nVidia chipset. Guest OS's are 32-bit FreeBSD (clock works fine after disabling ACPI, setting the clock source to the PIT, and running the guest tools), WinXP (unknown clock status), and i686 CentOS 5.2 (here is the problem). I've tried pretty much everything to try to fix it. I have host.cpukHz, host.noTSC, and ptsc.noTSC set in /etc/vmware/config. I've booted my kernel with noapic, nosmp, noacpi, divider=10. Sadly, I hit the clocksource=pit with divider bug, so I have not been able to boot with both that and divider=10, although clocksource=pit without a divider also does not work. I even built a custom kernel with SMP and APIC disabled, CPU_HZ=100, and booted with clocksource=pit noacpi, and it also gains time. Could anyone provide a recommendation as to what I can do to fix this problem? -- mouse, n: A device for pointing at the xterm in which you want to type. Confused by the strange files? I cryptographically sign my messages. For more information see <http://www.elehack.net/resources/gpg>.

3 6

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

virt