On Monday, March 21, 2011 at 10:32 AM, Ben De Luca wrote: Sorry to jump into the converstation, how slow can the deletion of > files actually be? > > One of the tests I ran a few weeks ago had me generating files, > deleting them and then writing them again from a number of clients. I > noticed that the space would never freed up again. I have my OSD's and > their journals on dedicated partions. > > I had planned on asking more on this once I had a stable system again.
Well I don't think the time is actually bounded, and it can be slowed down if multiple clients have capabilities on the file. The main issue is that the MDS is responsible for telling the OSDs to delete all the objects which make up the file; generally this shouldn't take too long but if you delete enough files quickly enough it might take some time. If you're watching it and the used space doesn't start dropping within ~20 minutes from last access then there's a problem, though. Given the scenario you describe it wouldn't surprise me too much if you just went quickly enough that the object deletions never actually occur (ie, they end up getting used again before the MDS has the permissions and the time to delete them). If faster space frees are actually a requirement for some reason you can comment on Bug #630 (http://tracker.newdream.net/issues/630). :) -Greg -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
