On Wed, Aug 09, 2017 at 01:13:02PM +0200, Michal Privoznik wrote:
On 08/09/2017 11:41 AM, Martin Kletzander wrote:
On Mon, Aug 07, 2017 at 02:20:06PM +0200, Michal Privoznik wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=1467245

Currently, there's a bug when undefining a domain with NVRAM
store. Basically, the unlink() of the NVRAM store file happens
during the undefine procedure iff domain is inactive. So, if
domain is running and undefine is called the file is left behind.
It won't be removed in the domain cleanup process either
(qemuProcessStop). One of the solutions is to remove if
regardless of the domain state and rely on qemu having the file
opened. This still has a downside that if the domain is defined
back the NVRAM store file is going to be new, any changes to the
current one are lost (just like with any other file that is
deleted while a process has it opened). But is it really a
downside?


It might be.  Why don't we disable removing it when the domain is
running?  We have some precedence for this.  The place which already
deals with this possibility is tools/virsh-domain.c in the cmdUndefine()
where we handle --remove-all-storage.  If you look at the help of that
command it also says:

 --nvram          remove nvram file, if inactive

And that makes sense to me.  What doesn't, on the other hand, is that it
also says:

 --keep-nvram     keep nvram file, if inactive

I don't get the "if inactive" there.  But I'm not going to check who
pushed that.  At least not again =)

Okay, I'll write a patch that:

a) forbids undefine for active domains, unless
b) KEEP_NVRAM flag is specified.


c) or there is no nvram =)

Yeah, well, that's caused by the unfortunate design of the way nvram is
handled by qemu and qemu driver.

Thanks

Michal

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

Attachment: signature.asc
Description: Digital signature

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

Reply via email to