We have a Sun thumper 34 terabyte, with 24T free. I''ve been asked to find out whether we can remove some disks from the zpool/ZRAID config (say about 10T) and install Veritas volumes on those, then migrate some data to it for block-level replication over a WAN. I know, horrifying - but the problem is there is no network/block-level replication inherent in (production Solaris) ZFS and we need it. I hate Veritas with a passion. But that''s another discussion :-) Thanks... -- This message posted from opensolaris.org
----- Original Message -----> From: "Forrest Aldrich" <forrie at gmail.com> > To: zfs-discuss at opensolaris.org > Sent: Monday, 24 May, 2010 6:47:40 PM > Subject: [zfs-discuss] Removing disks from a ZRAID config?> We have a Sun thumper 34 terabyte, with 24T free. I''ve been asked to > find out whether we can remove some disks from the zpool/ZRAID config > (say about 10T) and install Veritas volumes on those, then migrate > some data to it for block-level replication over a WAN. > > I know, horrifying - but the problem is there is no > network/block-level replication inherent in (production Solaris) ZFS > and we need it. I hate Veritas with a passion. But that''s another > discussion :-) > > Thanks...I dont know much about Veritas, but couldnt you install it on a zvol? -- iMx imx at streamvia.net
On May 24, 2010, at 10:47 AM, Forrest Aldrich wrote:> We have a Sun thumper 34 terabyte, with 24T free. I''ve been asked to find out whether we can remove some disks from the zpool/ZRAID config (say about 10T) and install Veritas volumes on those, then migrate some data to it for block-level replication over a WAN. > > I know, horrifying - but the problem is there is no network/block-level replication inherent in (production Solaris) ZFS and we need it.http://www.oracle.com/us/products/servers-storage/storage/storage-software/031764.htm -- richard -- ZFS and NexentaStor training, Rotterdam, July 13-15, 2010 http://nexenta-rotterdam.eventbrite.com/
> > Can you elaborate? > > > > Veritas has it''s own filesystem -- we need the block-level > > replication functionality to backup our data (live) over the WAN to > > a disaster > > recover location. Therefore, you wouldn''t be able to use Veritas > > with ZFS filesystem.zfs create -V 10G test/testvol newfs /dev/zvol/rdsk/test/testvol Or format the zvol however you need to for veritas, the above would create ufs, the key being you pass it the device rather than mount the file system; this would in theory then allow you to snapshot, locally, the volume etc. Check out the below article: http://www.markround.com/archives/37-ZFS-as-a-volume-manager.html "While browsing the ZFS man page recently, I made an interesting discovery: ZFS can export block devices from a zpool, which means you can separate "ZFS the volume manager" from "ZFS the filesystem". This may well be old news to many; however I haven''t seen many references to this on the web, so thought I''d post a quick blog update. " -- iMx imx at streamvia.net www.slashdevslashnull.com
I''ve seen this product mentioned before - the problem is, we use Veritas heavily on a public network and adding yet another software dependency would be a hard sell. :( -- This message posted from opensolaris.org
> > Thanks for the pointer, I will look into it. > > > > The first thing that comes to mind is a possible performance hit, > > somewhere with the VxFS code. I could be wrong, tho.No worries, certainly worth looking into though - if performance is acceptable, it could be a good solution. Let me know how it goes ;) -- iMx imx at streamvia.net www.slashdevslashnull.com
Forrest Aldrich wrote:> I''ve seen this product mentioned before - the problem is, we use > Veritas heavily on a public network and adding yet another software > dependency would be a hard sell. :(Be very certain that you need synchronous replication before you do this. For some ACID systems it really is required. And, sadly, it''s also required for rebootless client NFS failover (devs, will we ever get a zfs send/recv + nfs server that can preserver nfs file handles?) For most use cases async replication is just fine, and is far less fragile (block level replication happily replicates corruption). At $office we have 2 systems replicating on SRDF, everything else is async replication via NetApp SnapMirror or ZFS send/recv. To answer your original question, no, you can''t yet shrink a pool (it''s been coming Real Soon Now for a couple of years - no clue how much longer before it actually arrives). So you''d need enough temporary storage to replicate your data while you destroy and re-create your pool. -- Carson