[vbox-dev] split 2g vmdk seems to have some issues

Klaus Espenlaub Klaus.Espenlaub at Sun.COM
Mon Nov 2 16:12:48 GMT 2009


Huihong Luo wrote:
> thansk for the quick fixes
>  
> I verified that work well now.
>  
> yes, those two issues were caused by same bug

Bad news: the fix was incorrect, and it turned out it only worked around 
a bug elsewehere. Additionally the "fix" had the side effect of breaking 
raw partition VMDK files.

http://www.virtualbox.org/ticket/5355

Latest svn should have a better fix.

Note that all split VMDK files created by VirtualBox up to 3.0.10 are 
subtly broken (they incorrectly specify a non-zero sector offset for all 
extents but the first one).

It's possible to automatically repair such images, but this takes quite 
some effort, as the extent description is not touched by VirtualBox for 
existing images. I don't believe many people have split VMDK files 
created by VirtualBox, so I'm not sure if it's worth the effort.

Klaus

> --- On *Tue, 10/27/09, Klaus Espenlaub /<Klaus.Espenlaub at Sun.COM>/* wrote:
> 
> 
>     From: Klaus Espenlaub <Klaus.Espenlaub at Sun.COM>
>     Subject: Re: [vbox-dev] split 2g vmdk seems to have some issues
>     To: "Klaus Espenlaub" <Klaus.Espenlaub at Sun.COM>
>     Cc: "Huihong Luo" <huisinro at yahoo.com>, vbox-dev at virtualbox.org
>     Date: Tuesday, October 27, 2009, 5:01 AM
> 
>     Klaus Espenlaub wrote:
>      > Huihong Luo wrote:
>      >> for example
>      >>  (1) create a 40G disk as split2G, then attach to a xp guest vm,
>     it won't format
>      >
>      > Confirmed. Investigating the reason for the incorrect reporting
>     of VERR_OUT_OF_RANGE for perfectly valid sector numbers.
>      >
>      >> (2) VDCopy(VD_VMDK_IMAGE_FLAGS_SPLIT_2G) throws
>     VERR_OUT_OF_RANGE sometimes
>      >
>      > This is certainly caused by the same issue. However I don't see
>     any reason why this should be "sometimes", as the out of range check
>     is against data which isn't supposed to change after an image is
>     created.
> 
>     As suspected, there was just one bug which caused the misbehavior.
> 
>     Fixed in the OSE repository, and will be fixed as well in VirtualBox
>     3.0.10.
> 
>     Klaus
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> vbox-dev mailing list
> vbox-dev at virtualbox.org
> http://vbox.innotek.de/mailman/listinfo/vbox-dev




More information about the vbox-dev mailing list