Hi, a move across mount points should default to mv+rm, because these are 2 different file systems from the OS' point of view. However, a move within the same mount point should be atomic. Unfortunately, there is an exception: if the move crosses directories with quotas set. This had been fixed some time ago but, unfortunately, reverted back to bad behaviour. See https://tracker.ceph.com/issues/48203. If you are affected by this, please make some noise. The change applied is a serious performance regression working around a close to irrelevant corner case (it does not even solve it!) and it would be great if the devs would pick up on the suggestion to add a mount option "fast_move" to enable the usual atomic move on demand, with the low cost sanity checks applied. Best regards, ================= Frank Schilder AIT Risø Campus Bygning 109, rum S14 ________________________________________ From: Magnus HAGDORN <Magnus.Hagdorn@xxxxxxxx> Sent: 18 May 2022 15:45:54 To: ceph-users@xxxxxxx Subject: Re: Moving data between two mounts of the same CephFS Hi Mathias, I have noticed in the past the moving directories within the same mount point can take a very long time using the system mv command. I use a python script to archive old user directories by moving them to a different part of the filesystem which is not exposed to the users. I use the rename method of a python Path object which is atomic. In your case I would expect a copy and unlink because the operation is across different mount points. Is this operation a one off or a regular occurance? If it is a one off then I would do it as adminstrator. If it is a regular occurance I would look into re-arranging the filesystem layout to make this possible. Regards magnus On Wed, 2022-05-18 at 13:34 +0000, Kuhring, Mathias wrote: > This email was sent to you by someone outside the University. > You should only click on links or attachments if you are certain that > the email is genuine and the content is safe. > > Dear Ceph community, > > Let's say I want to make different sub-directories of my CephFS > separately available on a client system, > i.e. without exposing the parent directories (because it contains > other > sensitive data, for instance). > > I can simply mount specific different folders, as primitively > illustrated here: > > CephFS root: > - FolderA > - FolderB > - FolderC > > Client mounts: > - MountA --> cephfs:/FolderA > - MountB --> cephfs:/FolderB > > Now I'm wondering what actually happens in the background when I move > (not copy) data from MountA to MountB. > In particular, is CephFS by chance aware of this situation and > actually > performs an atomic move internally? > Or is more like a copy and unlink operation via the client? > > I appreciate your thoughts. > > Best wishes, > Mathias > > _______________________________________________ > ceph-users mailing list -- ceph-users@xxxxxxx > To unsubscribe send an email to ceph-users-leave@xxxxxxx The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th’ ann an Oilthigh Dhùn Èideann, clàraichte an Alba, àireamh clàraidh SC005336. _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |