virt

virt@lists.centos.org

1608 discussions

OpenNode (CentOS + OpenVZ + KVM) Management Server preview is now available
by Andres Toomsalu 15 Dec '10

15 Dec '10

Hi, We have been releasing first public preview of OpenNode Management Server and its ajax Web Management Console. OpenNode host (eg bare-metal OpenNode ISO installer) is basically a CentOS + OpenVZ + KVM + libvirt + func mashup with customized installer - while OpenNode Management Server provides central management features based on func and libvirt layer and including a RPC API server + ajax Web Management Console on top of func infrastructure - together with secure HTML5 VNC and SSH VM consoles. More info about OpenNode in general: http://opennode.activesys.org/ OpenNode Management Server release announcement: http://opennode.activesys.org/single/news/opennode-management-server-and-it… Regards, -- ---------------------------------------------- Andres Toomsalu, andres(a)active.ee CEO, Active Systems LLC

1 0

Storage performance
by Dennis Jacobfeuerborn 14 Dec '10

14 Dec '10

Hi, I recently did some benchmarking on a Rackspace VM and was surprised that bonnie++ showed a read throughput of almost 500MB/sec. Does anyone have an idea how they achieve these speeds in a shared environment? While you can achieve this with a RAID easily these days once you have lots of VMs accessing that array I would expect that speed to go down quite a bit. Was I just lucky that I was the only one accessing the disks that way at that moment or how do they keep up with this kind of demand? Regards, Dennis

3 3

VMs died due to hanging httpd processes
by Dennis Jacobfeuerborn 12 Dec '10

12 Dec '10

Hi, about an hour ago two web-serving VMs died at the same time with the following error on the console: INFO: task httpd:4304 blocked for more than 120 seconds. "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. httpd D 00af1f714d1112e2 0 4304 22471 4305 4303 (NOTLB) ffff88006574bdc8 0000000000000282 00000000000041f8 ffff88006574bea8 000000000000000a ffff88009747b820 ffffffff804f4b00 00000000001a5eee ffff88009747ba08 ffff880095be5015 Call Trace: [<ffffffff8022d03c>] mntput_no_expire+0x19/0x89 [<ffffffff8020eeae>] link_path_walk+0xa6/0xb2 [<ffffffff80263a7e>] __mutex_lock_slowpath+0x60/0x9b [<ffffffff80223f33>] __path_lookup_intent_open+0x56/0x97 [<ffffffff80263ac8>] .text.lock.mutex+0xf/0x14 [<ffffffff8021b52d>] open_namei+0xea/0x6d5 [<ffffffff8029cb30>] set_process_cpu_timer+0xc7/0xd2 [<ffffffff80227caa>] do_filp_open+0x1c/0x38 [<ffffffff8021a364>] do_sys_open+0x44/0xbe [<ffffffff802602f9>] tracesys+0xab/0xb6 Monitoring show that in a timeframe of about 3 minutes the load on the systems shot up to over 400 before they died. Since MaxClients is set to 512 I suspect that the processes had a mass-lockup with each process constantly causing a load of 1 (similar to what happens when a process hangs on an NFS mount point). One of the two VMs acts as a NFS server and exports directories to the other VM (but doesn't mount any external NFS sources itself). What is strange is that both system locked up at the same time since they are running on two different physical hosts. The hosts run Centos 5.3 while the VMs run Centos 5.5 as PV Xen guests. Since the call trace looks identical on both cases I wonder if anyone has an idea what exactly went wrong here? Regards, dennis

2 1

Re: [CentOS-virt] IBM GPFS filesystem
by Evan Fraser 03 Dec '10

03 Dec '10

Hi Adam, GPFS nodes can be licensed in two different ways. As a node with storage locally attached (called NSDs), or as a client node that connects to NSD servers. My Virtual servers are client nodes, and thus have an appropriate client license (these are cheap). They connect via GPFS's RDMA over Infiniband to my GPFS NSD servers which are a much more expensive license. The purpose of me doing it this way was to let me use storage already allocated to GPFS for my cluster without having to dedicate any luns to Virtual servers. Cheers, Evan. >Hi Evan, > >Thanks for the response. Just out of curiosity, do you have to pay the >extra licensing costs for each CentOS node? > >best, > >...adam This message and any attachments contain information that may be RMS Inc. confidential and/or privileged. If you are not the intended recipient (or authorized to receive for the intended recipient), and have received this message in error, any use, disclosure or distribution is strictly prohibited. If you have received this message in error, please notify the sender immediately by replying to the e-mail and permanently deleting the message from your computer and/or storage system.

1 0

IBM GPFS filesystem
by Adam Wead 03 Dec '10

03 Dec '10

Hi all, I was wondering if anyone might be able to speak about using IBM's GPFS filesystem as a means of storing virtual guests in a clustered environment with CentOS as the nodes and KVM as the hypervisor? I'm looking at using IBM's TSM software for archiving data from disk to tape. This requires buying a license for GPFS which is used in conjunction with TSM but can also be used as a clustered filesystem as well. As I understand it, GPFS can work with CentOS so long as you're using the right kernel. Is anyone out there using CentOS+GPFS for their virtualization environment? many thanks in advance, ...adam

10 15

PXDI; Automated Xen 4.0.1 + 2.6.32 dom0 installer for RHEL 6
by Digimer 22 Nov '10

22 Nov '10

I tool Pasi Kärkkäinen's tutorial[1] that he announced on Thursday and turned it into an automated installation program called PXDI. - http://wiki.alteeve.com/index.php/Pxdi I've tested it on ~9 machines and, so far, it has worked. Of course, this is a small sample set and could use much more testing. *hint* :) As it is built now, it can take a "minimal" Red Hat Enterprise Linux 6.0 x86_64 install (+perl) and completely built and install qemu, (patched) libvirt, xen 4.0.1 and a 2.6.32 dom0. It also adds itself to /boot/grub/grub.conf, so it should be a matter of run, wait, reboot, done. *PLEASE* don't run it on a server you care about. I have done my best to avoid errors, but I can make promises that it won't mess up your machine. Run it on test servers only, please, until more testing is done. If the installer aborts at any time, it is smart enough to (mostly) know what has already been done and will pick up where it left off. So it should be safe to run several times if needed without problem. Of course, I hope one run will be enough. :) Eventually we hope to have RPMs in repos so most users won't need this. For now though, I hope this will do. I would *love* to hear feedback; Good, bad or otherwise! Digimer 1. http://wiki.xen.org/xenwiki/RHEL6Xen4Tutorial -- Digimer E-Mail: digimer(a)alteeve.com AN!Whitepapers: http://alteeve.com Node Assassin: http://nodeassassin.org

1 0

How does KVM handle multiple cores?
by MargoAndTodd 22 Nov '10

22 Nov '10

Hi All, How does KVM handle multiple cores. I have an x5650 with 6 real cores that presents itself to my OS as 12 virtual cores (hyperthreading). Does KVM see 6 or 12 cores. And, can I tell KVM how many cores I want it to use? Am I misunderstanding how KVM works? What I am after is if my guest is 100% busy, I still want some power left over for my host. Many thanks, -T

3 4

KVM and Windows "/use pmtimer"
by Eric Searcy 20 Nov '10

20 Nov '10

http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/5/html/Virtualiz… Want to get some more people's opinion on this: the above doc says to use the boot parameter "/use pmtimer" to use the RTC "instead of the TSC for all time sources which resolves guest timing issues". One: does this have any bearing on whether the host has the constant_tsc flag (i.e. are all the sections that follow "Configuring hosts without a constant Time Stamp Counter" subordinate to it, or does that just have bearing on power management on that one CPU listed---AMD rev F)? Other question: http://support.microsoft.com/kb/938448 implies that for 2003 SP2 "/use pmtimer" shouldn't be needed as it will use the platform timer (RTC) if you have ACPI or APIC present. ("By default, Windows Server 2003 Service Pack 2 (SP2) uses the PM timer for all multiprocessor Advanced Programmable Interrupt Controller (APIC) Hardware Abstraction Layers (HALs) or Advanced Configuration and Power Interface (ACPI) HALs".) Anybody have any experience as to whether this (using ACPI feature in KVM) resolves the timing issues without needing pmtimer explicitly set? Thanks, Eric

2 1

What are KVM guest cores?
by MargoAndTodd 19 Nov '10

19 Nov '10

Hi All, What are guest cores in KVM? Are they fake, like everything else in the guest? Just another process running on the host emulating a core? Or are the guest cores actually connected directly to the physical cores on my motherboard? Many thanks, -T

4 8

what scheduling algorithm does KVM use?
by Nick 19 Nov '10

19 Nov '10

This may not be the best place to ask, but I was prompted by a question about guest cores on KVM. We currently use VMWare Server (v1.0) on CentOS5. It supports up to two virtual CPUs, but not very well, as I understand it. VMWare Server 2.0 might do better at supporting the same maximum of 2 CPUs, but if my research is correct, they both use what is called "strict co-scheduling". Which means, if a two virtual-CPU VM is waiting for a time slice on the physical host, there needs to be two physical CPUs available to be allocated. This is because amount of time scheduled to each virtual CPU has to stay equal, else assumptions in the virtualised software may become false and things start going wrong. The problem with this is that it is obviously more of a constraint to have two physical CPUs available. Therefore adding virtual CPUs to a co-scheduled VM can actually make performance worse if the physical CPUs are under any sort of contention. Performance degrades rapidly under load - which is exactly what you don't want for a webserver. Therefore advice seems to be to avoid using multi-virtual CPU VMs. [1] VMWare ESX uses something supposedly better, called relaxed co-scheduling. [2]. We're not likely to use that any time soon, since we are using 1&1 root servers running CentOS. I did try and research the algorithm(s) used by KVM. Apparently it doesn't use co-scheduling, but possibly something called the "completely fair scheduler", but I'm no an expert in VM scheduling and I didn't manage to discover what this means, nor specifically what the implications are for performance were in terms of scaling up the number of virtual CPUs on KVM. (I understand that the number of virtual CPUs supported by KVM is be quite high.) Anyway, I've been asked how to scale a VM up utilise more - say four or more - physical CPUs. Is KVM better at this than VMware Server, or does the same basic problem persist? Thanks, Nick 1. http://communities.vmware.com/thread/169323 http://cs.gmu.edu/~hfoxwell/cs671projects/southern_v12n.pdf > "The SMP > scheduling algorithms used by ESX server and by Xen > have a low overhead for single-threaded compute inten- > sive workloads. However, they do not scale as well for > multi-threaded workloads on a system with overcommit- > ted CPU resources." 2. http://communities.vmware.com/thread/157849 http://communities.vmware.com/docs/DOC-4960 http://communities.vmware.com/docs/DOC-5501

2 4

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

virt