[PVE-User] On VMware vSphere

Adam Thompson athompso at athompso.net
Thu Oct 2 19:37:10 CEST 2014


On 14-10-02 10:57 AM, Dietmar Maurer wrote:
>> PVE documentation is almost nonexistent compared to VMware, and what does
>> exist is of far lower quality - this is really the big deal, where "you get what you
>> pay for" with VMware.
> I also want to mention that the full source code is available... IMHO, this is the best
> Documentation you can get.

While that is true, it's also true that most users (i.e. sysadmins, in 
this case) don't have the time or the technical skill to read and 
understand the source code.
There are advantages to having the source code available, even if you 
don't have the ability to make local changes.  This is a partly 
technical, partly philosophical-and/or-political argument.

> I am quite sure that OpenVZ containers are less overhead that Xen. 

Yes, they would be, of course.
I didn't cover that since the OP was asking for comparison to VMware 
which has no similar feature.

Xen's DomU approach produces lower overhead than KVM, if you can live 
with the technical limitations (mainly: limited guest OS support).
OpenVZ has even less overhead than DomU, if you can live with the fairly 
restrictive technical limitations inherent in using Containers.

Kernel Same-Page Merging largely takes care of the overhead caused by 
running multiple copies of the same OS in multiple VMs.
Disk-based deduplication can manage the overhead caused by having 
multiple copes of the same OS on disk.
There's no way that I know of to avoid the processor and network 
overhead of having multiple copies of the OS running in parallel, which 
is what DomU partially accomplishes and OpenVZ fully accomplishes.

If containerization is supported for the OP's business applications, 
then OpenVZ provides a good way to minimize resource usage and increase 
performance.

-- 
-Adam Thompson
  athompso at athompso.net




More information about the pve-user mailing list