We are on 3.16 and Ceph 0.94.2, so I will try that parameter out and see how it goes. Thanks for the tip! On Tue, Jul 28, 2015 at 2:38 AM, Ilya Dryomov <idryomov@xxxxxxxxx> wrote: > > On Mon, Jul 27, 2015 at 11:38 PM, Wyllys Ingersoll > <wyllys.ingersoll@xxxxxxxxxxxxxx> wrote: > > I found a thread from 2012 talking about the maximum number of RBD > > devices that can be mapped, it seems to cap out at around 228. There > > was discussion about possibly fixing this by using a registered/fixed > > major block device and bumping the minor device number for each new > > mapping. > > > > http://www.spinics.net/lists/ceph-devel/msg09362.html > > > > Did this ever progress any further or was an official bug ever filed? > > Im seeing this issue now and was just wondering if there are > > workarounds (other than not creating that many mappings). > > What are your kernel and ceph versions? > > It's been fixed in 3.14 with the addition of single_major parameter to > rbd.ko. It's off by default because old userspace doesn't know how to > deal with it. If you are on 3.14 or later and your userspace is new > enough (0.75 or later IIRC), rbd map will load with single_major=Y. > > The new theoretical limit is 65536 rbd devices, but I must confess > I haven't tried to map more than ~4000. > > Thanks, > > Ilya -- 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
