Ah, crap. Okay, I think we should take the existing TMAPRM op and re-add the ENOENT check, and then change the mds to use a new TMAPRMSLOPPY op that doesn't error out. sage On Tue, 27 Nov 2012, Dan Mick wrote: > tmap_rm() no longer fails on nonexistent keys > (29fae494d0b1459c8bb934d42446e0ada7355402) > > causing rbd_remove of nonexistent image to succeed without error > (because, in the case that there's no header object, rbd_remove > keeps trying to clean up other bits like rbd_directory; in so doing, it relies > on tmap_rm() to fail removing the key from rbd_directory, or else it assumes > that this was an old-style image that we've just removed the last trace of and > returns success) > > This breaks test_rbd.test_remove_dne in test/pybind/test_rbd.py > > > We could change librbd obviously. Did you scan for other users, > though?...maybe there are more lurking > > Thoughts? > > > -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
