On Mon, Feb 22, 2010 at 01:09:52PM -0500, Laine Stump wrote:
FYI, I just verified that the restore failures I was seeing after applying this patch were actually happening *without* the patch as well, and are unrelated to domain save (it's a race condition in domain restore that needs to be dealt with separately), so this patch is okay to put in
I verified I've been testing with an unmodified form of this patch, *EXCEPT* that I hadn't done make syntax-check on it (since I didn't really think that it was working code at the time ;-)), and there is one occurence of white-space at the end of a line.
Should I resend with that change? Or do you want to just fix it up?
No, that's fine.
Also, notice that this patch saves the domain file with 0660 permission (umask will normally turn it into 0640) as we had thought that would be part of the way to allow restore from a root-squashed NFS server (just make sure that the reader had group read permissions). Now it seems we will be using the trick of running the restore code setuid instead, so the 0660 mode will no longer necessary. Should I revise this patch to create the file as 0600, or just do that as part of the upcoming domain restore patch?
0660 will be OK, since we setgid() too. Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://deltacloud.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|