dennis chen wrote: > Hi All, > > Recently I want to use IBM AIX as the ceph client to take use of the > cephFS and rbd service from ceph, > but after some investigation, seem there's no support for AIX as the > ceph client. Is there some clues about this? or is there some workaround > that I can map a block device created > from ceph cluster to the AIX system client? If you have a reasonably-recent Linux box and are using the v1 metadata format for RBD, you ought to be able to use the RBD kernel driver and re- export the RBD device over some variant of SCSI (iSCSI, SCSI over FC, etc). If you are using v2 metadata or don't have a recent Linux box, you may be able to run a QEMU virtual machine, using the QEMU RBD support to bring the disk into the VM and Linux's SCSI target code to export it as above. Another possible solution would be to use qemu-nbd to bypass needing a virtual machine, but I'm not at all sure of its performance/stability characteristics, nor whether AIX has the ability to access NBD devices. Note that all of these are almost certain to impair both resilience and performance compared to accessing it directly, but there really isn't much for AIX. As far as CephFS goes, there is a FUSE client, although I suspect AIX doesn't support FUSE. The only other option I can think of is to reexport it as another filesystem from another machine - there's ongoing efforts to support exporting CephFS via the Ganesha NFS server and the Samba SMB/CIFS server. -- 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
