Yes, technically the pool could be used for multiple things, but that is not a common practice and I don't think it is a scenario tested by the dev team with each release. I would about it unless you have a very compelling argument to do something so differently.
On Tue, Jul 25, 2017, 6:13 AM <[email protected]> wrote: > Understood. > > Would you recommend to have a dedicated pool for the data that is directly > written using librados and another pool for the filesystem (CephFS)? > > > > > 24. Juli 2017 19:46, "David Turner" <[email protected] > <%22david%20turner%22%20%[email protected]%3E>> schrieb: > > You might be able to read these objects using s3fs if you're using a > RadosGW. But like John mentioned, you cannot write them as objects into the > pool and read them as files from the filesystem. > On Mon, Jul 24, 2017, 12:07 PM John Spray <[email protected]> wrote: > > On Mon, Jul 24, 2017 at 4:52 PM, <[email protected]> wrote: > > Hello! > > > > I created CephFS according to documentation: > > $ ceph osd pool create hdb-backup <pg_num> > > $ ceph osd pool create hdb-backup_metadata <pg_num> > > $ ceph fs new <fs_name> <metadata> <data> > > > > I can mount this pool with user admin: > > ld4257:/etc/ceph # mount -t ceph 10.96.5.37,10.96.5.38,10.96.5.38:/ > /mnt/cephfs -o name=admin,secretfile=/etc/ceph/ceph.client.admin.key > > Need to untangle the terminology a bit. > > What you're mounting is a filesystem, the filesystem is storing it's > data in pools. Pools are a lower-level concept than filesystems. > > > ld4257:/etc/ceph # mount | grep ceph > > 10.96.5.37,10.96.5.38,10.96.5.38:/ on /mnt/cephfs type ceph > (rw,relatime,name=admin,secret=<hidden>,acl) > > > > To verify which pool is mounted, I checked this: > > ld4257:/etc/ceph # ceph osd lspools > > 0 rbd,1 templates,3 hdb-backup,4 hdb-backup_metadata, > > > > ld4257:/etc/ceph # cephfs /mnt/cephfs/ show_layout > > WARNING: This tool is deprecated. Use the layout.* xattrs to query and > modify layouts. > > layout.data_pool: 3 > > layout.object_size: 4194304 > > layout.stripe_unit: 4194304 > > layout.stripe_count: 1 > > > > So, I guess the correct pool "hdb-backup" is now mounted to /mnt/cephfs. > > > > Then I pushed some files in this pool. > > I think you mean that you put some objects into your pool. So at this > stage you have not created any files, cephfs doesn't know anything > about these objects. You would need to really create files (i.e. > write to your mount) to have files that exist in cephfs. > > > I can display the relevant objects now: > > ld4257:/etc/ceph # rados -p hdb-backup ls > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:7269 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:6357 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:772 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:14039 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:1803 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:5549 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:15797 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:20624 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:7322 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:5208 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:17479 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:14361 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:16963 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:4694 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:1391 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:1199 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:11359 > > MTY:file:8669fdbb88fda698afbac6374d826cba133a8d11:11995 > > [...] > > > > (This is just an extract, there are many more object.) > > > > Now, the question is: > > Can I display these files with CephFS? > > Unfortunately not -- you would need to write your data in as files > (via a cephfs mount) to read it back as files. > > John > > > > > When I check the content of /mnt/cephfs, there's only one directory > "MTY" that I have created; this directory is not related to the output of > rados at all: > > ld4257:/etc/ceph # ll /mnt/cephfs/ > > total 0 > > drwxr-xr-x 1 root root 0 Jul 24 15:57 MTY > > > > THX > > _______________________________________________ > > ceph-users mailing list > > [email protected] > > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > _______________________________________________ > ceph-users mailing list > [email protected] > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > > >
_______________________________________________ ceph-users mailing list [email protected] http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
