Den tis 16 juli 2019 kl 18:15 skrev Oscar Segarra <oscar.segarra@xxxxxxxxx>:
Hi Paul,That is the initial question, is it possible to recover my ceph cluster (docker based) if I loose all information stored in the etcd...I don't know if anyone has a clear answer to these questions..1.- I bootstrap a complete cluster mons, osds, mgr, mds, nfs, etc using etcd as a key store2.- There is an electric blackout and all nodes of my cluster goes down and all data in my etcd is lost (but muy osd disks have useful data)
Not having run container versions in my case made me reply wrongly last time, but I think it is still true that a running ceph will not store anything useful in etcd, but the info stuffed there by your deployment scripts is just _a_ way to get keys out to newly made containers while deploying, and not the _only_ way.
Also, a running ceph cluster will (probably!) not read anything useful out of etcd later on, since it keeps its own databases for all keys. Hence, the specific question about "I have ceph running, it was deployed using scripts and etcd for transport of keys, then it bombs out and for reason X, all etcd files are gone but rest is still there, then it boots up" should probably give you a working ceph cluster back, assuming no other part (like network configs, firewall rules or so) needs etcd for its own purposes. If the OS is gone, or all of /etc is gone or something like that, then the machine will not boot and ceph will not run but that is kind of obvious.
May the most significant bit of your life be positive.
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
