Dan, this sound weird: how can you run "cephfs /mnt/mycephfs set_layout 10" on a unmounted mountpoint? My client says: root@gw1:~# cephfs /mnt/ceph/ set_layout -p 3 Error setting layout: Inappropriate ioctl for device And in IRC I've found that "path need to be a path to an already mounted cephfs" (http://irclogs.ceph.widodh.nl/index.php?date=2012-11-26) -- Marco Aroldi 2013/3/15 Dan van der Ster <dan@xxxxxxxxxxxxxx>: > We eventually resolved the problem by doing "ceph mds add_data_pool > 10; cephfs /mnt/mycephfs set_layout 10", where 10 is the id of our new > "data" volume, and then rebooting the client machine (since the cephfs > mount was hung). > Cheers, Dan > > On Fri, Mar 15, 2013 at 4:13 PM, Marco Aroldi <marco.aroldi@xxxxxxxxx> wrote: >> Same here, >> Now mounting cephfs hangs for a minute then says "mount error 5 = >> Input/output error" >> Since the new pool has id=3, I've also executed "ceph mds >> add_data_pool 3" and "ceph mds remove_data_pool 0" >> The monitor log has this line: >> >> 2013-03-15 16:08:08.327049 7fe957441700 0 -- 192.168.21.11:6789/0 >> >> 192.168.21.10:0/491826119 pipe(0x1b94c80 sd=23 :6789 s=0 pgs=0 cs=0 >> l=0).accept peer addr is really 192.168.21.10:0/491826119 (socket is >> 192.168.21.10:54670/0) >> >> -- >> Marco Aroldi >> >> 2013/3/15 Dan van der Ster <dan@xxxxxxxxxxxxxx>: >>> Hi, >>> >>> >>> On Fri, Mar 15, 2013 at 9:52 AM, Sebastien Han <sebastien.han@xxxxxxxxxxxx> >>> wrote: >>>> >>>> Hi, >>>> >>>> It's not recommended to use this command yet. >>>> >>>> As a workaround you can do: >>>> >>>> $ ceph osd pool create <my-new-pool> <pg_num> >>>> $ rados cppool <my-old-pool> <my-new-pool> >>>> $ ceph osd pool delete <my-old-pool> >>>> $ ceph osd pool rename <my-new-pool> <my-old-pool> >>> >>> >>> >>> We've just done exactly this on the default pool data, and it leaves cephfs >>> mounts in a hanging state. Is that expected? >>> >>> Cheers, Dan >>> >>>> >>>> >>>> >>>> –––– >>>> Sébastien Han >>>> Cloud Engineer >>>> >>>> "Always give 100%. Unless you're giving blood." >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> PHONE : +33 (0)1 49 70 99 72 – MOBILE : +33 (0)6 52 84 44 70 >>>> EMAIL : sebastien.han@xxxxxxxxxxxx – SKYPE : han.sbastien >>>> ADDRESS : 10, rue de la Victoire – 75009 Paris >>>> WEB : www.enovance.com – TWITTER : @enovance >>>> >>>> On Mar 15, 2013, at 9:27 AM, Marco Aroldi <marco.aroldi@xxxxxxxxx> wrote: >>>> >>>> Hi, >>>> >>>> I have a new cluster with no data. >>>> Now it has 44 osd and my goal is to increase in the next months to >>>> reach a total of 88 osd. >>>> >>>> My pgmap is: >>>> pgmap v841: 8640 pgs: 8640 active+clean; 8730 bytes data, 1733 MB >>>> used, 81489 GB / 81491 GB avail >>>> 2880 PG each for data, metadata and rbd pools >>>> This value was set by mkcephfs >>>> >>>> Chatting on IRC channel, it was told me to calculate 100 pg per Osd >>>> and round to near power-of-2 >>>> So in my case would be 8192 PG for each pool, right? >>>> >>>> My question: >>>> Knowing to have to double the number of osd, >>>> is it advisable to increase pg_num right now with the following commands? >>>> ceph osd pool set data pg_num 8192 --allow-experimental-feature >>>> ceph osd pool set metadata pg_num 8192 --allow-experimental-feature >>>> ceph osd pool set rbd pg_num 8192 --allow-experimental-feature >>>> >>>> Thanks >>>> >>>> -- >>>> Marco Aroldi >>>> _______________________________________________ >>>> ceph-users mailing list >>>> ceph-users@xxxxxxxxxxxxxx >>>> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com >>>> >>>> >>>> >>>> _______________________________________________ >>>> ceph-users mailing list >>>> ceph-users@xxxxxxxxxxxxxx >>>> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com >>>> >>> _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |