>From my experience you’ll be better off planning exactly how many OSD’s and nodes you’re going to have and if possible equip them from the start. By just adding a new drive to the same pool ceph will start to rearrange data across the whole cluster which might lead to less client IO depending on what you’re comfortable with. In a worst case scenario, your clients won’t have enough IO and your services might be ”down” until it’s healthy again. Rebuilding 60 x 6TB drives will take quite some time. Each SATA drive has about 75MB-125MB throughput at best, so a rebuild of once such drive would take approx. 16-17 hours. Usually it takes some x2 or x3 times longer in a normal case and if your controllers or network is limited. // david > 21 mars 2016 kl. 13:13 skrev Bastian Rosner <bro@xxxxxxxx>: > > Yes, rebuild in case of a whole chassis failure is indeed an issue. That depends on how the failure domain looks like. > > I'm currently thinking of initially not running fully equipped nodes. > Let's say four of these machines with 60x 6TB drives each, so only loaded 2/3. > That's raw 1440TB distributed over eight OSD nodes. > Each individual OSD-node would therefore host "only" 30 OSDs but still allow for fast expansion. > > Usually delivery and installation of a bunch of HDDs is much faster than servers. > > I really wonder how easy it is to add additional disks and whether chance for node- or even chassis-failure increases. > > Cheers, Bastian > > Am 2016-03-21 10:33, schrieb David: >> Sounds like you’ll have a field day waiting for rebuild in case of a >> node failure or an upgrade of the crush map ;) >> David >>> 21 mars 2016 kl. 09:55 skrev Bastian Rosner <bro@xxxxxxxx>: >>> Hi, >>> any chance that somebody here already got hands on Dell DSS 7000 machines? >>> 4U chassis containing 90x 3.5" drives and 2x dual-socket server sleds (DSS7500). Sounds ideal for high capacity and density clusters, since each of the server-sleds would run 45 drives, which I believe is a suitable number of OSDs per node. >>> When searching for this model there's not much detailed information out there. >>> Sadly I could not find a review from somebody who actually owns a bunch of them and runs a decent PB-size cluster with it. > > _______________________________________________ > 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
 |