virt March 2014

virt@lists.centos.org

23 participants
17 discussions

MS-Win7 kvm guest gets dhcp from host bridge
by James B. Byrne 23 Jun '14

23 Jun '14

CentOS-6.5 i86_64 qemu-kvm-0.12.1.2-2.415.el6_5.4 MS-Windows v7proSP1 We have installed a MicroSoft Win7 system as a guest and have joined it to our MS AD domain. The system runs and has internet access. However, the IPv4 address it obtains and the gateway it is assigned are both sourced from the virtual machine host system and not from our AD DC DHCP server. To clarify, the virbr0 IP address is 192.168.122.1 and the Windows guest has that address as its gateway and an IP address in the 192.168.122.0/24 netblock. This is not as we desire as the internal addresses we regularly assign to MS workstations via DHCP are supposed to belong to a different netblock entirely. This permit filtering on incoming and outgoing traffic at the gateway. I realize that there is a DHCP service running on virbr0 for the purpose of provisioning guests with system control traffic but I do not know how to limit that to its intended purpose and allow another DHCP server to provide the IP address to the windows guests. I hope that problem description is not too confusing. Can anyone provide me with some guidance on the matter? Are kvm guests required to have either static ip addresses or dhcp addresses provided by the host system? Thanks in advance. -- *** E-Mail is NOT a SECURE channel *** James B. Byrne mailto:ByrneJB@Harte-Lyne.ca Harte & Lyne Limited http://www.harte-lyne.ca 9 Brockley Drive vox: +1 905 561 1241 Hamilton, Ontario fax: +1 905 561 0757 Canada L8E 3C3

4 3

Xen Security patching
by Chris Elliott 07 Jun '14

07 Jun '14

Hi Guys When is the next update of Hypervisor packages going to be released? There appears to be no changes to the main hypervisor RPMs since September (Even in the BETA / RC1 tree) and there are 5+ Xen Security Advisories with patches which need to be added. "Security issues are handled and released in sync with public disclosure upstream with zero lag" I can rebuild my own, but that's not really the point. In other news I've been doing some testing of the 3.10 kernel and so far so good! Thanks - Chris

3 7

VMs failing to restart
by Lars Hecking 09 May '14

09 May '14

This is a problem I've had on and off under CentOS5 and CentOS6, with both xen and kvm. Currently, it happens consistently with kvm on 6.5, e.g. with every kernel update. I *think* it generally worked fine with the 6.4 kernels. There are 7 VMs running on a 6.5, x86_64, 8GB RAM host, each with 512MB RAM and using the e1000 NIC. I picked this specific NIC because the default does not allow reliable monitoring through SNMP (IIRC). The host has two bonded NICs with br0 running on top. When the host reboots, the VMs will generally hang bringing up the virtual NIC, and I need to go through several iterations of destroy/create, for each VM, to get them running. The always hang here (copy&paste from console): ... Welcome to CentOS Starting udev: udev: starting version 147 piix4_smbus 0000:00:01.3: SMBus Host Controller at 0xb100, revision 0 e1000: Intel(R) PRO/1000 Network Driver - version 7.3.21-k8-NAPI e1000: Copyright (c) 1999-2006 Intel Corporation. ACPI: PCI Interrupt Link [LNKC] enabled at IRQ 11 e1000 0000:00:03.0: PCI INT A -> Link[LNKC] -> GSI 11 (level, high) -> IRQ 11 e1000 0000:00:03.0: eth0: (PCI:33MHz:32-bit) 00:16:3e:52:e3:0b e1000 0000:00:03.0: eth0: Intel(R) PRO/1000 Network Connection Any suggestions on where to start looking?

4 9

AWS Images updated for new instance types
by Karanbir Singh 22 Apr '14

22 Apr '14

hi, · CentOS 6.3 (x86_64) – Release Media: https://aws.amazon.com/marketplace/pp/B00A6L6F9I · CentOS 6.4 (x86_64) – Release Media: https://aws.amazon.com/marketplace/pp/B00DGYP804 are now marked for and updated to run with the m3 and c3 instance types. I'm still working with the amazon crew to get the 6.5 images online and an automated pipeline proposed ( hopefully, also adopted ) to get this image injection speeded up in the future. Regards -- Karanbir Singh +44-207-0999389 | http://www.karan.org/ | twitter.com/kbsingh GnuPG Key : http://www.karan.org/publickey.asc

2 6

injecting a key into the xen images
by Karanbir Singh 02 Apr '14

02 Apr '14

hi, As a part of the test suite for xen that I've started off - I needed a way to inject a ssh key into the image [1]; so have come up with this : https://github.com/CentOS/sig-virt-t_xen/blob/master/scripts/inject_ssh.sh ; its not pretty and it wont handle lots of use cases, but it does what is needed at hand. Comments ? I'm also considering if we should include this script within the .bz2 file for the image itself ( and for all such images ) Further feature creep, should we now extend this script a bit to also include a 'download and inject ssh key' with some param's like : ./get_image.sh 6.5 x86_64 <path to sshkey> and that can hit http://cloud.centos.org/<calculated path to image>/ ; download it, and add the key. This script could live within centos-release-xen dramatically reducing the barrier to entry and get-go with the xen4centos stack. thoughts ? - KB 1: qcow2, http://cloud.centos.org/centos/6/devel/CentOS-6-x86_64-xen-pv.bz2 -- Karanbir Singh +44-207-0999389 | http://www.karan.org/ | twitter.com/kbsingh GnuPG Key : http://www.karan.org/publickey.asc

4 6

Xen updates / TODO
by Pasi Kärkkäinen 25 Mar '14

25 Mar '14

Hello, A couple of TODO items/suggestions: 1) Xen 4.2.4 was just released: http://lists.xen.org/archives/html/xen-devel/2014-02/msg01624.html I think we should first update Xen4Centos6 rpms to Xen 4.2.4 to have a "known good" release from the stable branch currently in use in Xen4Centos6. 2) After that, we should start packaging/testing Xen 4.3.2: http://lists.xen.org/archives/html/xen-devel/2014-02/msg01622.html Xen 4.3 brings much improved XL toolstack in addition to a lot of other changes/fixes. 3) After updating to Xen 4.3.2 rpms start troubleshooting/debugging the libvirt libxl driver and get it working with Xen4Centos6 system.. How does that sound like? :) -- Pasi

6 34

Re: [CentOS-virt] centos5 xen domu with X: screens not found [SOLVED]
by Peter Peltonen 20 Mar '14

20 Mar '14

Hi, On Tue, Mar 18, 2014 at 3:55 PM, Stefano Stabellini < stefano.stabellini(a)eu.citrix.com> wrote: > > Strangely enough according to xenstore, vfb is running and connected. > > Can you try connecting to it using xm or vncviewer directly? > > Or maybe setup the guest to use a simple fbcon just to try out if it > works, and then setup the more complicated environment you are looking > for afterwards? Thank you Stefano for your help. Just to be sure I did the virtualization again (used dd to copy the machine's system and then edited Xen stuff in place) and - surprise, surprise - X login window appeared without any problems in virt-manager. So problem solved. Best, Peter

1 0

Problems with Windows on KVM machine
by engineer＠colocat.ru 20 Mar '14

20 Mar '14

Hello. We've started a virtualisation project and got stuck in one moment. Currently we are using the following: Intel 2312WPQJR as a node Intel R2312GL4GS as a storage with Intel Infiniband 2 ports controller Infiniband Mellanox SwitchX IS5023 for commutation. The nodes run CentOS 6.5 with built-in Infiniband package (Linux v0002 2.6.32-431.el6.x86_64), the storage - CentOS 6.4 also built-in drivers (Linux stor1.colocat.ru 2.6.32-279.el6.x86_64). On the storage is made an array, that is shown in system as /storage/s01. Then it is exported via NFS. The nodes connect to NFS by: /bin/mount -t nfs -o rdma,port=20049,rw,hard,timeo=600,retrans=5,async,nfsvers=3,intr 192.168.1.1:/storage/s01 /home/storage/sata/01 mount shows: 192.168.1.1:/storage/s01 on /home/storage/sata/01 type nfs (rw,rdma,port=20049,hard,timeo=600,retrans=5,nfsvers=3,intr,addr=192.168.1.1) Then we create a virtual machine with virsh with a disk bus virtio. All is OK, until we don't start Windows on KVM. It may work for 2 hours or 2 days, but under heavy load it hangs the mount (i.e. /sata/02 and 03 are accessible, but requesting 01 will result in a total hang of console). This can be beaten only by hardware reset of the node. If we mount without rdma - all is fine. What can we do in such case? If any debugs or info needed - please ask. Best regards, Nikolay.

3 8

centos5 xen domu with X: screens not found
by Peter Peltonen 18 Mar '14

18 Mar '14

Hi, I am in a process to virtualize an existing non-virtualized CentOS5 system. Everything else works fine except X: virt-manager's console just hangs after bootup. I can use the text console though and what I see in X server's log is the error message "screens not found". The only graphics related setup I've done is in the domu config this line: vfb = [ "type=vnc,vncunused=1,keymap=fi" ] Is there something else I should configure on the guest or in dom0? Regards, Peter

2 3

Fwd: Xen4CentOS kernel panic on dom0 reboot
by Karl Johnson 14 Mar '14

14 Mar '14

I don't think it's related to microcode_ctl as my server was rebooting just fine before switching to Xen4CentOS in December 2013, then the reboot issue started. Note that it's not only on Dell servers but also Supermicro so I don't think it's hardware related too. Karl On Thu, Mar 13, 2014 at 8:14 PM, PJ Welsh <pjwelsh(a)gmail.com> wrote: > Anyone think it's related to microcode_ctl as noted in > http://lists.us.dell.com/pipermail/linux-poweredge/2013-October/048538.html? > I will see about testing for this issue ASAP... > > pjwelsh > > > On Thu, Mar 13, 2014 at 6:58 PM, PJ Welsh <pjwelsh(a)gmail.com> wrote: > >> Comments at bottom: >> >> >> On Tue, Mar 11, 2014 at 12:32 PM, Phillippe Welsh <pjwelsh(a)gmail.com>wrote: >> >>> Comments inline: >>> >>> ----- Original Message ----- >>> > From: "Pasi Kärkkäinen" <pasik(a)iki.fi> >>> > To: "Discussion about the virtualization on CentOS" < >>> centos-virt(a)centos.org> >>> > Sent: Sunday, March 9, 2014 4:32:13 AM >>> > Subject: Re: [CentOS-virt] Fwd: Xen4CentOS kernel panic on dom0 reboot >>> > >>> > On Sat, Mar 08, 2014 at 09:09:07AM -0600, PJ Welsh wrote: >>> > > No, I have not followed those instructions yet. These were >>> > > production >>> > > servers that I had scheduled firmware updates late Sunday >>> > > evening. The >>> > > first time I though the error was a fluke and only began to >>> > > research it >>> > > after the second failure (and still no firmware updates due to >>> > > the >>> > > power-cycle). I may try to sneak in a restart of one of the >>> > > systems late >>> > > Sunday night US CT. >>> > > >>> > >>> > OK. >>> >>> I ran the "stop" for all of the xen related pieces in the order that the >>> /etc/rc3.d/ had them. >>> The VM's did not shutdown and the /usr/lib64/xen/bin/qemu-dm STUFF >>> entries were left behind running. >>> Since I could not xm shutdown any longer, I killed off all qemu-dm >>> proccesses and attempted a reboot... >>> HUNG on the reboot with the prepended umount error messages... >>> >>> > >>> > > Still not sure why the running vm's would stop the reboot... The >>> > > server >>> > > shows that it was suppose to be restarting. I have had a similar >>> > > stuck on >>> > > restarting message (minus all the umount errors) on some Dell >>> > > T105's >>> > > running CentOS 6.5 and the "reboot=pci" grub.conf kernel option >>> > > is what >>> > > ended up working for them. I have not tested that possible >>> > > option yet, >>> > > either since that would take 2 reboots to put into place. >>> > > >>> > >>> > Yeah, it's worth testing both, to figure out what's wrong. >>> >>> Next reboot attempt included the "reboot=pci" grub.conf kernel option... >>> No affect :( >>> HUNG on the reboot with the prepended umount error messages... >>> >>> I ran out of time to attempt an xm shutdown for each VM manually, then >>> reboot. >>> >>> What's interesting is that when I do an lsof on the file system that is >>> unable to umount, the *only* connected PID's are the qemu-dm ones, but not >>> *all* of them.??? >>> >>> Thanks >>> >>> PJ >>> ... >>> >> >> UPDATE: I cleanly shut down *all* vm's and unmounted the filesystem that >> had the umount issue noted previously and then issued the reboot command. >> *STILL* the Dell R710 will be hung at the rebooting line. >> >> No reboot possible on 2 Dell R710's with at least the 2 most recent >> CentOSXen4 kernels. >> >> Any other suggestions? >> >> Thanks >> >> pjwelsh >> > >

2 1

Xen4CentOS installation strangeness
by Peter Peltonen 14 Mar '14

14 Mar '14

Hi, I have a server with Supermicro X7DVL-3 (P9) motherboard, 16G ECC RAM and LSI SAS 1068e RAID controller. I installed CentOS 6.5 64bit on the machine without any problems, but after following the Xen setup steps at http://wiki.centos.org/HowTos/Xen/Xen4QuickStart which installed me the kernel 3.10.32-11.el6.centos.alt.x86_64, I encountered a problem: After "Starting certmonger [OK]" the screen went black and the system became unresponsive: keyboard was not working (NumLock did not respond) and SSH was not responding either. After first lockup I increased dom0 max mem to 2G, but rebooting after that produced the same result. The strange thing is, that after a third reboot everything worked ok: screen went black for a moment after "Staring certmonger [OK]" but after that the graphical login screen appeared and I could use the system normally. The fourth reboot went ok as well. Any ideas what could cause this kind of behaviour? Regards, Peter

4 4

Fwd: Xen4CentOS kernel panic on dom0 reboot
by Phillippe Welsh 14 Mar '14

14 Mar '14

Subject: Re: [CentOS-virt] Xen4CentOS kernel panic on dom0 reboot On Wed, Mar 5, 2014 at 10:17 AM, David Vrabel <david.vrabel at citrix.com>wrote: > > On 05/03/14 15:09, Karl Johnson wrote: > > > I've been using Xen4CentOS for the last 3 months. It's working fine and > > > dom0/domUs are stable but the server does a kernel panic when doing a > > > reboot and the server has to be hard reset manually. It has kernel panic > > on the 3 last reboot. > > > > There is a xenbus device still present and during shutdown it is trying > > to set it to CLOSED but at this point xenstored isn't running and the > > xenbus write stalls. > > > > Do you have VMs that are still running when you attempt a reboot? If so > > shutting them down will likely avoid this. > > > > Can you provide the output of xenstore-ls prior to attempting a reboot? > > > > > > > I though Xen init.d scripts would stop all of them before rebooting? Here's > the output of chkconfig and xenstore-ls: > > http://pastebin.centos.org/8186/ > > Thanks, > > Karl I've got the "me-too" on the reboot hang issue for 2 different Dell R710's with xen-4.2.4-29.el6 and at least the kernels kernel-3.10.25-11.el6.centos.alt.x86_64 and kernel-3.10.23-11.el6.centos.alt.x86_64. I have not tried to reboot with the latest kernel-3.10.32-11.el6.centos.alt.x86_64 (if kernel even makes a difference). I *have* had dom0_mem=1024M,max:1024M option in place for all of them with only 6 VM's. Any new suggestions? pjwelsh

3 6

Re: [CentOS-virt] [CentOS] KVM and DHCP
by James B. Byrne 13 Mar '14

13 Mar '14

On Thu, March 13, 2014 15:49, Reindl Harald wrote: > > > Am 13.03.2014 20:43, schrieb James B. Byrne: >> CentOS-6.5 >> >> We have a KVM guest running MS-WinV7pro. This guest is joined to an Active >> Directory Domain. That domain provides DHCP to the members. The KVM guest >> does not obtain its IP from the domain but from the local host's qemu >> hypervisor instead. >> >> Is there anyway to get around this and have the guest MS-Win OS get its DHCP >> from the same place as the rest of the domain members? > > smells like you are using NAT for your guest instead > bridge it to the host network, the guest does not > know from where it gets DHCP infos and there should > not be more than one dhcpd on a network segment, anything > else leads to race-conditions - the faster dhcpd wins > > http://www.dedoimedo.com/computers/kvm-bridged.html > > maybe the article is outdated, try google as i did > https://www.google.at/search?q=kvm+nat+versus+bridging > > Yes, that was it. Apparently if one creates a kvm guest via the virt-manager gui application then NAT is automatically selected as the default. It may be that the other guests were originally hand-crafted via virsh and the br0 interface was specified for them, but I doubt that. I have been using virt-manager since day one on this project and I am getting a queasy feeling that virt-manager's behaviour has obscurely changed in significant ways since I first began using it. This is not the first problem that I have had creating new guests with recent versions of virt-manager which were not in evidence with earlier use. In any case, to fix this via virt-manager's __virtual hardware details__ interface one must select the __NIC__ and change the settings from <Virtual network default 'NAT'> to <Specify shared device name> and then input the bridge interface name <br0> in the text box that subsequently appears. One must also change the Device model: from <Hypervisor default> to <virtio>. Thanks for the pointer. -- *** E-Mail is NOT a SECURE channel *** James B. Byrne mailto:ByrneJB@Harte-Lyne.ca Harte & Lyne Limited http://www.harte-lyne.ca 9 Brockley Drive vox: +1 905 561 1241 Hamilton, Ontario fax: +1 905 561 0757 Canada L8E 3C3

1 0

KVM and DHCP
by James B. Byrne 13 Mar '14

13 Mar '14

CentOS-6.5 We have a KVM guest running MS-WinV7pro. This guest is joined to an Active Directory Domain. That domain provides DHCP to the members. The KVM guest does not obtain its IP from the domain but from the local host's qemu hypervisor instead. Is there anyway to get around this and have the guest MS-Win OS get its DHCP from the same place as the rest of the domain members? -- *** E-Mail is NOT a SECURE channel *** James B. Byrne mailto:ByrneJB@Harte-Lyne.ca Harte & Lyne Limited http://www.harte-lyne.ca 9 Brockley Drive vox: +1 905 561 1241 Hamilton, Ontario fax: +1 905 561 0757 Canada L8E 3C3

2 1

Are the "centos.org" AMIs in the EC2 marketplace "official"? locale breakage in 6.4 AMI
by Dan Pritts 13 Mar '14

13 Mar '14

Hi all, I've reviewed the mailing list archives and it looks like the "centos.org" AMIs in the marketplace are really official products of the centos project (whatever that means), but I wanted to ask explicitly to confirm. Presuming yes, I've run into an odd problem with the 6.4 x86_64 AMI (not updated). with a few packages, files are not being properly created in /usr/share/locale/xx/LC_XXXX when I install the RPM. I mostly don't care about the files themselves, but I worry about what else is wrong. I'll use subversion as my example below, but there are about half a dozen affected packages, including the coreutils package as shipped in the AMI (ami-bf5021d6). I have turned off selinux enforcement. RPM database shows the file: [root@dns-east svn]# rpm -ql subversion | grep de.LC /usr/share/locale/de/LC_MESSAGES/subversion.mo verify doesn't complain: [root@dns-east svn]# rpm --verify subversion [ just returns, no output ] RPM package contains the file: [root@dns-east svn]# rpm2cpio /tmp/subversion-1.6.11-10.el6_5.x86_64.rpm | cpio -t | grep de.LC_MESSAGES 23744 blocks ./usr/share/locale/de/LC_MESSAGES/subversion.mo [root@dns-east svn]# rpm2cpio /tmp/subversion-1.6.11-10.el6_5.x86_64.rpm | cpio --extract --make-directories 23744 blocks [root@dns-east svn]# ls -l ./usr/share/locale/de/LC_MESSAGES/ total 308 -rw-r--r--. 1 root root 312655 Mar 13 14:03 subversion.mo Uninstalling and reinstalling the RPM doesn't help. There are no crazy install scripts, either. [root@dns-east svn]# rpm -qp --scripts ../subversion-1.6.11-10.el6_5.x86_64.rpm postinstall scriptlet (using /bin/sh): # Register the snvserve service /sbin/chkconfig --add svnserve 2>&1 /sbin/ldconfig preuninstall scriptlet (using /bin/sh): if [ $1 = 0 ]; then /sbin/service svnserve stop > /dev/null 2>&1 || : /sbin/chkconfig --del svnserve > /dev/null 2>&1 || : fi postuninstall program: /sbin/ldconfig [root@dns-east svn]# -- Dan Pritts ICPSR Computing & Network Services University of Michigan +1 (734)615-7362

1 1

Remove Centos from AWS marketplace
by JacobV 09 Mar '14

09 Mar '14

https://forums.aws.amazon.com/thread.jspa?messageID=481859&#481859 https://forums.aws.amazon.com/thread.jspa?messageID=453572&#453572 This is a timebomb waiting to strike so many people who like do daily snapshot backups and keep them for few weeks and not realizing their snapshots are useless if they had accidentally mess up some boot related file earlier on. Another scenario you mess up the sudoers file or the root authorized keys - you'd have to loose a whole days data and go to previous nights restore just for single file error like that ? If AWS marketplace is unable to remove this hardcoded rule then it's only prudent to remove Centos from AWS marketplace and release it in the community section instead ? Thoughts ?

4 6

Xen4CentOS kernel panic on dom0 reboot
by Karl Johnson 05 Mar '14

05 Mar '14

Hello, I've been using Xen4CentOS for the last 3 months. It's working fine and dom0/domUs are stable but the server does a kernel panic when doing a reboot and the server has to be hard reset manually. It has kernel panic on the 3 last reboot. Please stand by while rebooting the system... INFO: task reboot:19800 blocked for more than 120 seconds "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message Call trace ... Picture of the call trace: http://i59.tinypic.com/169olt1.png CentOS: 6.5 Kernel: 3.10.23-11.el6.centos.alt.x86_64 Xen: 4.2.3 Server: Supermicro 6017R-WRF Any idea how to troubleshoot or fix this issue? We can't reboot the dom0 during the week because domUs (2) are in production and this server is in a datacenter. Kind regards, Karl

2 2

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

virt March 2014