[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Terribad "file:" performance (vs reasonable "phy:" perf)

On Feb 3, 2012, at 2:49 PM, Florian Heigl wrote:

> I don't have sufficient testing experience, but I do have experience
> concerning loop (randomly used it a few years from Xen2 to Xen3). Do
> NOT USE loop if you have any valuable data in the virtual machines.
> There is enough written about it in the archives here and most Xen books.

Thanks for this info.  Do you happen to have some URLs I could look at (or 
suggest a search term for me) to show that /dev/loopN is bad?  From my own 
testing, I would imagine it might have to do with sync semantics; I'd be very 
interested in seeing what the specific concerns are.

There seem to be a few alternatives to loop:

        1: blktap
        2: qemu-dm
        3: phy:/dev/sdXY    (blkback)
        4: LVM              (also blkback)

1: I don't believe I have access to the blktap driver.  For anyone else looking 
at disk performance issues and stumbles onto this message, here is my setup 
(which will probably not resemble yours):

> Specifically, my configuration is Xen-4.1.2 on a 64-bit Linux-3.1.0 dom0 
> (dom0_mem=2048), with a 64-bit Linux-3.1.0 domU configured as a PV host.  
> Neither host nor guest is multilib or has 32-bit libraries.  And, I'm using 
> xl (not xm) to start my instance.

2: Performance seems abysmal.

3: Yes, this is a potential last resort, but a fixed-partition scheme is truly 
horrid from a management perspective.

4: A new avenue to investigate.


Is there an option I'm unaware of...?


Xen-users mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.