Created attachment 1372339 [details] output with debug flags turned on Description of problem: guestmount -d centinil -i --ro mnt worked in fedora 26 to mount a read only snapshot of the centinil disk image on "mnt". On fedora 27 it fails and tells me to re-run the command with a lot of debug flags. (I'll attach the output from doing that). I see one of the debug output errors is: qemu-img: /tmp/libguestfssn6HDK/overlay1.qcow2: Failed to get shared "write" lock Is another process using the image? And the answer is, "yes" which is why I used --ro (which the man page says should work even on running VMs). Version-Release number of selected component (if applicable): libguestfs-tools-c-1.37.35-1.fc27.x86_64 How reproducible: 100% Steps to Reproduce: 1.see above 2. 3. Actual results: Error from guestmount Expected results: Mounted read only snapshot Additional info: guestmount works fine if I shutdown the VM before trying the mount. P.S. What the heck does "Erlang bindings for libguestfs" mean as a description for the libguestfs component. Isn't that a bit self-referential?
This is caused by a bug in libvirt, therefore I'm making this a duplicate of the other bug. You can work around this (temporarily, hopefully) using the direct backend: export LIBGUESTFS_BACKEND=direct Note that accessing a qcow2 image, even read only, isn't guaranteed to give sensible results. Your application needs to deal with failures. > P.S. What the heck does "Erlang bindings for libguestfs" mean as a > description for the libguestfs component. Isn't that a bit self- > referential? It seems to describe the purpose of the erlang-libguestfs package. What's the problem? *** This bug has been marked as a duplicate of bug 1513447 ***
... accessing a *live* qcow2 image ...
(In reply to Richard W.M. Jones from comment #1) > > P.S. What the heck does "Erlang bindings for libguestfs" mean as a > > description for the libguestfs component. Isn't that a bit self- > > referential? > > It seems to describe the purpose of the erlang-libguestfs package. > What's the problem? It is the description I get for the package "libguestfs" when I was adding this bugzilla (not the "erlang-libguestfs" package).
I do have the libvirt-3.7.0-3.fc27 mentioned in the other bug. If there is something else needed to get the --ro mount to work, it ought to be mentioned in the guestmount man page.
(In reply to Tom Horsley from comment #3) > (In reply to Richard W.M. Jones from comment #1) > > > > P.S. What the heck does "Erlang bindings for libguestfs" mean as a > > > description for the libguestfs component. Isn't that a bit self- > > > referential? > > > > It seems to describe the purpose of the erlang-libguestfs package. > > What's the problem? > > It is the description I get for the package "libguestfs" when I was adding > this bugzilla (not the "erlang-libguestfs" package). Hmm, you're not wrong about that. I've filed a bug: https://bugzilla.redhat.com/show_bug.cgi?id=1529142 (In reply to Tom Horsley from comment #4) > I do have the libvirt-3.7.0-3.fc27 mentioned in the other bug. If there is > something else needed to get the --ro mount to work, it ought to be > mentioned in the guestmount man page. Maybe I didn't get the right duplicate, but this is still a bug in libvirt. It's not a bug in libguestfs so it doesn't require changes to this package so far. Please follow up on the other bug.