Hello! I''m looking good cluster fs for a while and I''m very sad that I didn''t found any solution with 100% failproof. Only one I know now its googleFS, but its closed project, used by google, not open source. I''m ,ooking for a cluster file system with replicating - in simple example: 10 nodes, each with hard disk. Each stripe is residing always on 2 or more node. So I can power off any 1 node on any time, and clients have 100% stability - there is no error/outage, 100% SPOF free. Can it be done with Lustre? -- Best regards, Darek All your base belong to Me.
Brent A Nelson
2007-Feb-07 17:55 UTC
[Lustre-discuss] Lustre vs googleFS (high availability)
I, too, have been on a long quest for the holy grail of 100% uptime. Yes, Lustre is suited for this better than most. Lustre itself won''t do the replication, though, although they are planning just such a feature in the future. You can use Lustre combined with drbd (for network mirroring) or a SAN and some heartbeating software (such as Heartbeat) for failover of the services. However, you would have at least some sort of delay during the failover; it would not be instantaneous. I don''t know how quick you can make it. Thanks, Brent Nelson Director of Computing Dept. of Physics University of Florida On Thu, 8 Feb 2007, Darek wrote:> Hello! > > I''m looking good cluster fs for a while and I''m very sad that I didn''t > found any solution with 100% failproof. Only one I know now its > googleFS, but its closed project, used by google, not open source. I''m > ,ooking for a cluster file system with replicating - in simple > example: > 10 nodes, each with hard disk. Each stripe is residing always on 2 or > more node. So I can power off any 1 node on any time, and clients have > 100% stability - there is no error/outage, 100% SPOF free. Can it be > done with Lustre? > > -- > Best regards, > Darek > All your base belong to Me. > > _______________________________________________ > Lustre-discuss mailing list > Lustre-discuss@clusterfs.com > https://mail.clusterfs.com/mailman/listinfo/lustre-discuss > >
Aaron Knister
2007-Feb-10 08:31 UTC
[Lustre-discuss] Lustre vs googleFS (high availability)
Sounds like this might do the trick-- http://datafarm.apgrid.org/document/ -Aaron Brent A Nelson wrote:> I, too, have been on a long quest for the holy grail of 100% uptime. > > Yes, Lustre is suited for this better than most. Lustre itself won''t > do the replication, though, although they are planning just such a > feature in the future. You can use Lustre combined with drbd (for > network mirroring) or a SAN and some heartbeating software (such as > Heartbeat) for failover of the services. > > However, you would have at least some sort of delay during the > failover; it would not be instantaneous. I don''t know how quick you > can make it. > > Thanks, > > Brent Nelson > Director of Computing > Dept. of Physics > University of Florida > > On Thu, 8 Feb 2007, Darek wrote: > >> Hello! >> >> I''m looking good cluster fs for a while and I''m very sad that I didn''t >> found any solution with 100% failproof. Only one I know now its >> googleFS, but its closed project, used by google, not open source. I''m >> ,ooking for a cluster file system with replicating - in simple >> example: >> 10 nodes, each with hard disk. Each stripe is residing always on 2 or >> more node. So I can power off any 1 node on any time, and clients have >> 100% stability - there is no error/outage, 100% SPOF free. Can it be >> done with Lustre? >> >> -- >> Best regards, >> Darek >> All your base belong to Me. >> >> _______________________________________________ >> Lustre-discuss mailing list >> Lustre-discuss@clusterfs.com >> https://mail.clusterfs.com/mailman/listinfo/lustre-discuss >> >> > > _______________________________________________ > Lustre-discuss mailing list > Lustre-discuss@clusterfs.com > https://mail.clusterfs.com/mailman/listinfo/lustre-discuss > >-- "Computers are incredibly fast, accurate and stupid; humans are incredibly slow, inaccurate and brilliant; together they are powerful beyond imagination." --Albert Einsten Aaron Knister Center for Research on Environment and Water 4041 Powder Mill Road, Suite 302; Calverton MD 20705 Office: (240) 247-1456 Fax: (301) 595-9790 http://crew.iges.org -------------- next part -------------- A non-text attachment was scrubbed... Name: aaron.vcf Type: text/x-vcard Size: 313 bytes Desc: not available Url : http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20070210/59592749/aaron.vcf