Virtualisierung für Linux, Windows und Solaris Thomas Finnern DESY/IT-SysOps

IT-Seminar Agenda  Eine Einführung in die Thematik  Vergleich von XEN und Solaris Containern  Ein bisschen Windows …  Stand und Ausblick für unser Rechenzentrum  Live-Vorführung von Christoph und Martin

IT-Seminar Allgemeiner Einstieg

IT-Seminar Overview  Virtualization  Why  Features  How  Hypervisor  Guests  Use Cases  Where to  Conclusions

IT-Seminar Virtualization  Definition  Network Resources  Storage Resources  Used Virtualizations  Clustering Sun Cluster  Content Based Routing F5 Layer 7 Routing Cisco ?  Host Virtualization Solaris Container XEN  Linux  Windows Virtualization is the technique of managing and presenting storage devices and other resources functionally, regardless of their physical layout or location. In computing, paravirtualization is a virtualization technique that presents the abstraction of virtual machines with a software interface that is similar but not identical to that of the underlying hardware. Hypervisor OS Root OS User App

IT-Seminar Why  Minimize Efforts  Easy Provisioning  Easy Resource Control  Multi-OS Service Offer  Security / Service Separation  Getting Better  Enhanced Load Distribution  Enhancing Fault Tolerance and Security  Separate Test, Developing and Production on same Hardware

IT-Seminar How  Sun  Solaris 10 Zones  Plan: (Open)Solaris 11 with XEN  ZFS (Solaris 10 U3+)  Zone + SRM = Container  RedHat  Start: RHEL 5.0 with XEN 3.0  Plan: RHEL 4.5+, 5.0+ with Built-In Support including AFS  Logical Volume Manager  Hypervisor and Guests

IT-Seminar Key Features and Benefits  XEN:  Open Source  Complete Separation between Guests  Paravirtualization  Intel-Vanderpool and AMD-Pacifica Supported for unmodified Guests (without Live Migration)  Virtual Machines near Native Speed  Live Migration of Guests  Up to 32 virtual CPU’s per Guest  Load Distribution and Weighting between Guests  X86/32 and x86/64 Support  Network over Software Bridge(s)  Own Mac-Address for each Interface  Solaris Container:  Open Solaris  Separation (and Shares) configurable  One OS for All  Sparc and x86  Guests Machines at Native Speed  No Live Migration (?, cloning planned)  Cores per Zone not limited  Resource Management between Containers  Sparc and x86 with 32/64 bit Support  Network over selected Device(s)  No own Mac-Address for each Device (?)  Sparse-root and Whole-root Models

IT-Seminar  Customisation  Image Preparation Partitioning OS and Data Application Networking  Deployment Methods mk_image mk_virtual  Live Cycle Management As Usual … Hypervisor / Global Zone  XEN Test:  Defined one admin domain cross platform  Configuration: Add feature vms  Reconfiguration: New Install / Reboot  Hypervisor with Minimum Installation  Logical Volume Manager for FS-Handling  Separate Network for Service and Production  Guest Installations from Scratch/Image ?  Unvisible Filesystems /scratch/vsges/root /scratch/vsges/sge /scratch/vsges/…  Changing to Image Preparation  Solaris Container Test:  Defined one admin domain cross platform  Configuration: Add feature vms  Reconfiguration: On the fly  Global Zone Installation will/can be inherited  ZFS for FS-Handling  Separate Network for Service and Production  Guest Installations from Scratch/Image  Visible Filesysstems /scratch/vinnetou/root /scratch/vinnetou/afscache /scratch/vinnetou/…  Methods  mk_image Install File or Clone/Copy Pack  mk_virtual Unpack Networking Partitioning Glue

IT-Seminar Guests / Zones and Pools  XEN Test:  Configuration: Add feature vms-  Openafs kernel module must be compiled  XEN-Mac 0x00FFFFFF&IP|0x (Change to XEN mac=00:16:3e:xx:xx:xx ?)  Solaris Container Test:  Configuration: Add feature vms-  afs can be inherited (ro ?)  Next version: privilege grant for kernel module usage  Slight differences in SVC (Service Management Facility) for own sshd startup  Testing NIS-Setup

IT-Seminar Expected Use Cases  Separation of Applications and Services  Providing Service IP Addresses  Application Specific User Registry and System Settings  License Cost Optimisation  e.g. Oracle “Capped Containers”  Application Specific Resource Binding  Storage  Network  CPU  Peak Load Elimination  Multi OS Offer for Customers  Test Environments  Developing Environments  Migration Support  Multi Customer Pools e.g. SGE  Automatic Deployment of e.g. WEB- Services  Educational Environments for Admins and Students  Database Factory  Security by Encapsulation  New Service Environments  Order and Delivery within 1 Hour  Central Image Handling over SAN  Ease Computer Adminstration  New Operating Concepts Installation Testing Monitoring Backup Patching ...

IT-Seminar Where to  Server Consolidation (Cost, Heat, Admin)  Running Solaris and Linux on one Host  Running 100-Core Hosts over Internal Crossbar as Farm or WEB Appliance  Cross Cluster Hopping  Live Migration  Image Moving  Configuration and Life Cycle Management  Buy xx Open Xen Licences

IT-Seminar Conclusions  Virtualization Is Simple To Use  Nice Operating and Support Model  Administrate only Big Boxes  Cross Platform Support  Future  Real or Para Virtualization will be standard  XEN Hypervisor for Solaris and vice versa  Standard provisioning frame (kernel, partition, image, data, …)  Security  Don’t let your system be virtualized by somebody else …

IT-Seminar  IT-SysOps:  WBOOM-Configuration Guest knows Hypervisor(-Cluster): SF=vms- Hypervisor knows all guests: woom:VMS  BIRD Cluster: Common Store in AFS Queue and Shadow Master (it-services01/2) 2 Host Login Cluster(it-services01/2) 16 Exec Cores (bird009/bird010)  Library spires-Container (spires)  WOF Test Apache (vinnetou,vinnern)  Plan: Netra-Consolidierung Netapp for Live-Migration  Yves Kemp:  Im Tier2-Batch “noch nicht”  1 OS SL3 -> SL4  Sven Sternberger:  Seit ca. 2 Jahren haben wir mit der Whithouse einen Server der den Entwicklern schnell neue SL3 bzw SL4 Installationen zur Verfuegung stellt. Sie habe die Moeglichkeit hier neue Software zu installieren und verschiedene Versionen von Installationen zu sichern. Durch die Nutzung von rsync ist es moeglich in wenigen Minuten zwischen verschiedenen Setups zu wechseln, und so die komplexen Setups aufzubauen die fuer eine Softwareentwicklung von einem verteilten System wie dCache notwendig ist.  Owen Synge:  Seit ca. 1 Jahr entwickelt Owen an einem System zur Automatisierung der build und test Umgebung. Das bedeutet ein Entwickler checkt eine neue Version ein in der Versionsverwaltung. Daraufhin wird automatisch eine frische SL3 bzw. SL4 Installation generiert, die build Umgebung wird installiert und aus den neuen Sourcen ein aktuelles Packet gebaut. Im folgenden werden mehrere Instanzen generiert, die mit der neuen Software ausgestattet werden und mehrere Testsequenzen werden abgefahren um die Funktionsfaehigkeit des neuen Releases sicherzustellen.

IT-Seminar Live Demo   Christopher Open Solaris  zoneadm list -vc  zlogin GUEST  ps -efZ  reboot  Martin Red XEN Linux  xm list  xm console GUEST  xm top  reboot

IT-Seminar Notizen