So in fact, there is no other solution than normal file-based backups and hoping the MDS won't crash? Are there any plans which release will contain a stable CephFS/MDS?
Well, if you've made changes to your data which impacted the metadata, and then you restore to a backup of the metadata pool, but not the data, then what's there isn't what CephFS thinks is there. Which would be confusing for all the same reasons that it is in a local filesystem. You could construct a use case and a backup system which would produce a working system, but I don't think it'd be very useful for most applications. -Greg Software Engineer #42 @ http://inktank.com | http://ceph.com On Wed, Apr 10, 2013 at 12:13 PM, Maik Kulbe wrote: > I think going backwards in time would be, what a backup is for, isn't it? ;) > > My question really just is, if it is possible, to back it up. It just really > stinks to completely rebuild the cluster and re-import all the data > everytime I make some small mistake in the environment and the metadata > backup would really help to save some time here. > >> If you were to do that you'd be going backwards in time with your >> metadata, so — not really. CephFS is not generally production-ready >> at >> this time, but we welcome bug reports! >> -Greg >> Software Engineer #42 @ http://inktank.com | http://ceph.com >> >> >> On Mon, Apr 8, 2013 at 12:52 PM, Maik Kulbe >> wrote: >> > Hi, >> > >> > I'm currently testing Ceph with the POSIX fs at work as a fs cluster. >> So far >> > I've managed to cripple the test environment three or four times be >> > unintendently crashing the MDS(1 active, 2 hot-standby). It seems to >> me the >> > MDS are pretty sensitive to all kinds of environmental changes and so >> to use >> > this I really need a backup solution for the Metadata. >> > >> > I've seen that RADOS supports an export command and that this also >> works >> > with the MD pool. Now the question is - can I just export and in case >> I need >> > it re-import the exported metadata or are there problems with that? If >> so, >> > what would be a better way to handle this? >> > >> > Thanks in advance, >> > >> > Maik Kulbe >> > _______________________________________________ >> > 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
