Hello Dave,
you can configure Ceph to pick multiple OSDs per Host and therefore work
like a classic raid.
It will cause a downtime whenever you have to do maintenance on a system,
but when you plan to grow it quite fast, it's maybe an option for you.
--
Martin Verges
Managing director
Hint: Secure one of the last slots in the upcoming 4-day Ceph Intensive
Training at
https://croit.io/training/4-days-ceph-in-depth-training.
Mobile: +49 174 9335695
E-Mail: martin.verges(a)croit.io
Chat:
https://t.me/MartinVerges
croit GmbH, Freseniusstr. 31h, 81247 Munich
CEO: Martin Verges - VAT-ID: DE310638492
Com. register: Amtsgericht Munich HRB 231263
Web:
https://croit.io
YouTube:
https://goo.gl/PGE1Bx
Am So., 2. Feb. 2020 um 05:11 Uhr schrieb Dave Hall <kdhall(a)binghamton.edu>du>:
Hello.
Thanks to advice from bauen1 I now have OSDs on Debian/Nautilus and have
been able to move on to MDS and CephFS. Also, looking around in the
Dashboard I noticed the options for Crush Failure Domain and further
that it's possible to select 'OSD'.
As I mentioned earlier our cluster is fairly small at this point (3
hosts, 24 OSDs) , but we want to get as much usable storage as possible
until we can get more nodes. SInce the nodes are brand new we are
probably more concerned about disk failures than about node failures for
the next few months.
If I interpret Crush Failure Domain = OSD, this means it's possible to
create pools that behave somewhat similar to RAID 6 - something like 8 +
2 except dispersed across multiple nodes. With the pool spread around
like this loosing any one disk shouldn't put the cluster into read-only
mode - if a disk did fail, would the cluster re-balance and reconstruct
the lost data until the failed OSD was replaced.
Does this make sense? Or is it just wishful thinking.
Thanks.
-Dave
--
Dave Hall
Binghamton University
_______________________________________________
ceph-users mailing list -- ceph-users(a)ceph.io
To unsubscribe send an email to ceph-users-leave(a)ceph.io