QuicksearchDisclaimerThe individual owning this blog works for Oracle in Germany. The opinions expressed here are his own, are not necessarily reviewed in advance by anyone but the individual author, and neither Oracle nor any other party necessarily agrees with them.
|
Is hypervisor based virtualisation the correct way?Monday, April 14. 2008Trackbacks
Trackback specific URI for this entry
No Trackbacks
Comments
Display comments as
(Linear | Threaded)
The killer-feature of hypervisors is live migration ... and maybe running different OS patchlevels.
Otherwise hypervisors waste too much resources (at least for I/O)
I donīt think about Live Migration as a killer feature. Live Migration creates his own class of problems like uncontinious time and on really loaded systems live migration isnīt an easy nor short task.
hypervisors like vmware and virtual box are cool for testing and such, but for production? no way. zones, jails and such is what i use.
things that require live migration should be clustered anyways. hardware and firmware level support needs to evolve for useful virtualization.
Are you referering to zones, of which best practice of patching is to delete and recreate them?
Yeah, itīs one of the best practices. Better than patching every vm on itīs own... and as the patching problem is solvable (or will disappear as such a problem, when the patchin process speeds up), the problem of having to patch each VM on itīs own isnīt solvable by the nature of this virtualisation.
The problem: Most people think of zones, as they are a static thing, but unlike VMs you can create a zones within seconds. You do not install a OS, you install an application execution environment. As soon youīve seperated the os and application it is more efficient to simply throw away a zone and installing a new one.
yes, that sucks. but if you look at nexenta, you see that it is possible to patch them easy and fast. i dont know if ips is the solution for solaris or if the patching issue will be fixed without it. from the technical view, it should not be that big problem to speed up patching with the current system.
what i did is isolating app data from the os and scripting the setup of the zones.
Pretty much both is happening. But we canīt change much of the old patching because of the compatibility guarantees for Solaris 10 and later.
IPS will solve that nicely for the Opensolaris reference distribution from sun. And by the way: Since the last version of Solaris you can liveupgrade Solaris 10 with non-global zones. The tip with deleting/creating zones is pretty much from the time when this feature wasnīt available. |
+1The LKSF bookThe book with the consolidated Less known Solaris Tutorials is available for download here
Web 2.0Contact
Networking xing.com My photos Buttons![]() This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 Germany License
![]() ![]() ![]() Blog AdministrationDonateOkay, okay ... as several people have asked for it ... but you know my opinion.
|