Ceph

I have hosts like cephdisk01 cephdisk02 cephdisk03 in my test cluster. I tried to name all my OSD's like osd0101...osd0313 (each host has 14 disks, so 14 OSD's). Then I noticed in the watchlog some messages like osd313, and using "ceph osd out 0313" also results in the zero being stripped (responds "osd313 not found").

This is annoying as zero padding helps things sort properly and lets everything be a fixed length. I am not sure if this also happens to hostnames (as I've zero-padded those too, but seems like they at least communicate) or mon/mds names, but if it does I also vote that those allow zero padding as well.

I would attempt a patch myself but was not sure if this was an implementation thing and shouldn't be changed for some reason.