Hi I have 2 questions on use of ZFS. How do I ensure I have site redundancy using zfs pools, as I see it we only ensures mirrors between 2 disks. I have 2 HDS on one each site and I want to be able to loose the one of them and my pools should still be running. F.inst. I have created 2 luns on each site (A and B) named AA, AB and BA, BB. I then create my pool and mirror AA to BA and AB to BB. If I lose site B hosting BA and BB can I be sure they do no hold both copies of any data? The second question is if there is any work going into providing shared pool support as I would love to see it on RAC, something like the QFS shared disk group support is what I am looking for? -- // Flemming Danielsen -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20061206/96eae7f7/attachment.html>
You specify the mirroring configuration. The top-level "vdevs" are implicitly striped. So if you, for instance, request something like zpool create mirror AA BA mirror AB BB then you will have a pool consisting of a stripe of two mirrors. Each mirror will have one copy of its data at each location. Hence if site B goes down, you''ll lose BA and BB, and each mirror will transition to a degraded state. So the answer to your first question is yes, with this configuration you can be sure that no data will reside solely on LUNs at site B. On your second question, I don''t believe there is work currently being done on shared pools, but I''ll let the ZFS team clarify the road map. This message posted from opensolaris.org
Le 06/12/2006 ? 05:05:55+0100, Flemming Danielsen a ?crit> Hi > I have 2 questions on use of ZFS. > How do I ensure I have site redundancy using zfs pools, as I see it we only > ensures mirrors between 2 disks. I have 2 HDS on one each site and I want to be > able to loose the one of them and my pools should still be running. F.inst. > > I have created 2 luns on each site (A and B) named AA, AB and BA, BB. I then > create my pool and mirror AA to BA and AB to BB. If I lose site B hosting BA > and BB can I be sure they do no hold both copies of any data? >I''m asking a question (maybe stupid) what you use to attach two 2disks on 2 different site ? You using FC attachement ? Regards. -- Albert SHIH Universite de Paris 7 (Denis DIDEROT) U.F.R. de Mathematiques. 7 i?me ?tage, plateau D, bureau 10 Heure local/Local time: Wed Dec 6 09:06:30 CET 2006
Hi We allocate in lun sizes of 60 Gb so to get to 120 gb, but this is not limited to 2 disks could be in the 100''s of luns. We currently plan putting somthing in the region of 20 to 30 TB in the pools using 60 gb lun sizes. But we will need share support at some point due to costumer demand of using RAC. On 06/12/06, Albert Shih <shih at math.jussieu.fr> wrote:> > Le 06/12/2006 ? 05:05:55+0100, Flemming Danielsen a ?crit > > Hi > > I have 2 questions on use of ZFS. > > How do I ensure I have site redundancy using zfs pools, as I see it we > only > > ensures mirrors between 2 disks. I have 2 HDS on one each site and I > want to be > > able to loose the one of them and my pools should still be running. > F.inst. > > > > I have created 2 luns on each site (A and B) named AA, AB and BA, BB. I > then > > create my pool and mirror AA to BA and AB to BB. If I lose site B > hosting BA > > and BB can I be sure they do no hold both copies of any data? > > > I''m asking a question (maybe stupid) what you use to attach two 2disks on > 2 > different site ? You using FC attachement ? > > Regards. > > > -- > Albert SHIH > Universite de Paris 7 (Denis DIDEROT) > U.F.R. de Mathematiques. > 7 i?me ?tage, plateau D, bureau 10 > Heure local/Local time: > Wed Dec 6 09:06:30 CET 2006 >-- // Flemming Danielsen -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20061206/1da69e14/attachment.html>