Limit on the number of virtual machines running on RHES / CentOS XEN?
Linux - ServerThis forum is for the discussion of Linux Software used in a server related context.
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Introduction to Linux - A Hands on Guide
This guide was created as an overview of the Linux Operating System, geared toward new users as an exploration tour and getting started guide, with exercises at the end of each chapter.
For more advanced trainees it can be a desktop reference, and a collection of the base knowledge needed to proceed with system and network administration. This book contains many real life examples derived from the author's experience as a Linux system and network administrator, trainer and consultant. They hope these examples will help you to get a better understanding of the Linux system and that you feel encouraged to try out things on your own.
Click Here to receive this Complete Guide absolutely free.
Host crashes after 4th virtual machines is started on RHES / CentOS XEN?
We've been experiencing sudden host server crashes minutes after starting a fourth virtual machine.
Our setup looks like:
Dell Poweredge T300
1 x Intel Xeon X3323 Quad Core 2.5 ghz
16 GB Ram
CentOS 5.3 (64 bit)
(we've got two of them... same problem on both)
Server is running a stripped down version of CentOS 5.3 (64-bit), running only the built-in Xen Virtualization Environment. There is no other services running on the server (not samba, httpd, sendmail, cups... nothing except Xen)
We've created several virtual machines, and as long as we don't start a fourth virtual machine everything runs smoothly (impresive hardware).
Each virtual server is configured as:
1 Virtual CPU
1 GB RAM
However, 5 minutes or so after starting a fourth virtual machine, the entire host server crashes and restarts itself.
Are we limited by the number of cores on the host machine CPU (4 cores)? 1 for the host and 3 for virtual machines?
We've read in forums about other Xen setups running up to 11 virtual machines on less powerful hardware? (a dual core server).
Should we be using FULLY VIRTUALIZED virtual machines instead?
Is the number of XEN virtual machines in fact limited by the number of cores? If so, how can someone run several virtual machines on a single core host?
Has anyone tried VirtualBox (Sun) or VMWare? (free alternatives)?
By the way, we were replacing a previous Dell Server (Poweredge 2600 with 512 MB Ram and a single Xeon single core processor running Open Virtuozzo). We were able to run up to 16 virtual machines at the same time. Of course none of the machines endured hard work (testing environments, etc). But hey, my point is that we expected to get a much higher number of virtual machines on this new hardware.
Anyone with a clue is welcome to comment.
Last edited by accend; 05-02-2009 at 10:21 PM.
Reason: Better insight into real problem
There is no limit at all, and never would be. With a paravirt system (which you do want, not fully virt - that would be a massive backwards step) you need physical ram to be assigned, but not CPU's. Would the behaviour you're experiencing make any sense if it was by design? Deliberately causing a system to crash?? Somehow I don't think so. I find my centos 5.3 Xen system spontaneously reboots at non-specific points in it's use, but as it's only a test lab it seldom matters enough to troubleshoot. Your's on the otherhand would preumably benefit with some exploration as to the reasons for the reboot. If it's as reproducible as you make out then there's plenty of scope to run more debugging and such.
No there's no limit, and if there was, do you really thing spontaneously nuking yourself is ever going to be a proper course of action??? if your scenario are as reproducible as you think, then you should have a good amount of scope to do some debugging and checking for known issues. Does this behvaiour vary across xen kernel versions, for example?
Would the behaviour you're experiencing make any sense if it was by design?
Exactly, that's what i thought. I'm not sure, but to my understanding CentOS has an unlimited virtual guest capability builtin, not unlike its twin brother RedHat Enterprise Server. RHES is limited to 4 virtual guests. Only RHES Advanced has unlimited virtual guests capability builtin. You can check their specs here: http://www.redhat.com/rhel/server/compare/
Even so, we're talking about 4 virtual guests.. not three. So i don't believe this has anything to do with our crashes.
Originally Posted by acid_kewpie
do you really thing spontaneously nuking yourself is ever going to be a proper course of action???
Obviously not. We have experienced this a few times just to confirm that the issue was reproducible. Virtual machines were holding no "real" apps or data. However we do not believe nuking ourselves is the proper course of action. Just explaining the situation.
Anyway, we've tried out a couple of versions on the xen-kernel. Same story.
We currently are using:
[root@srv10 ~]# uname -a
Linux srv10.accend.com.mx 2.6.18-128.1.6.el5xen #1 SMP Web Apr 1 09:53:14 EDT 2009 x86_64 x86_64 GNU/Linux
Logs aren't that insightfull; system instantly crashed and we haven't been able to find any useful info in them. Any suggestions on where to look for info that might be helpful?
Anyone else has used VirtualBox?
We need to run aprox 12 virtual machines on this server:
All of them will be used very "lightly". All of them are intended for web app development and testing among a group of 6 programers so no heavy duty use on any of them. (did we mention earlier that we are replacing a single core server with 512 ram running 10 virtual servers on open virtuozzo?)
Sorry to tell you this. But we were unable to solve the issue. We tried on 5 different machines (from quad core xeons down to pentium 4). We also tried CentOS and Red Hat OS.
But the good news is, we ended up using VirtualBox. It turned out to be an even more robust solution.
We're even migrating everything from VMWare to Virtualbox.
- Great GUI, same one for Mac, Linux, Win, Solaris.
- We can move a VM one host to another without any trouble.
- Backups and Snapshots are so easy
- Free. (we would gladly pay for this... really)
- Supports a hell lot of guest OSs
- Takes care of all the different tweaks usually requiered to run a virtual machine.
- Easiest installation of all Virtualization solutions, and upgrades are very easy to perform (usually on a 4 week cycle)
Our production environment runs on RedHat. Most of the development environment is using CentOS. VMWare was used on both prod and testing env.
We're using Virtualbox now in a lot of our infrastructure. We're even using it now on client's infrastructure.
VMWare support can solve a few things, but won't make things work if the techonology doesn't support it. Don't get me wrong, VMWare is great. However, we've found Virtualbox to be a better match to our needs. Community support is growing like hell and we've been able to solve everything.