Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Sep 17, 2009, at 8:33 PM, Gianluca Cecchi wrote: > On Thu, Sep 17, 2009 at 11:42 PM, Michael Tokarev <mjt at tls.msk.ru> > wrote: > Gianluca Cecchi wrote: > [] > > And what about instead do somthing like this: > > - sda and sdb on both nodes > - raid0 on both nodes so that on eac node you have one md0 device > - only one drbd resource based on md0 on both > > This is wrong. If either disk fails, whole raid0 fails, > ie, half the thing fails. > > With reverse approach, ie, two drbd resources on top of > disks, and raid0 on top of drbd resources - if any disk > fails only that 1/4 of whole thing fails. > > It's classical raid10. It's always done like > disk => mirroring => striping, but never like > disk => striping => mirroring. > > > - use drbd0 device as PV for your VG > - add clvmd to your cluster layer (you need cman too for clvmd) > > I'm doing it but only with one disk per node > > With one disk per node it's simple raid1. > > /mjt > > > > My approach is done in reality and is named raid0+1, while the > initial post one is socalled raid1+0 (aka raid10 as you wrote) > > See also http://en.wikipedia.org/wiki/Nested_RAID_levels > > BTW: in general, failures to take care of happens not only at hw > level (hard disk in your comment) but also in sw (in the scenario > here, possibly in sw raid layer or drbd layer, for example) > > Gianluca I'm not so concerned with raid1+0 vs raid0+1 but the use of striping on the same volume between two active nodes. Any pros/cons for md or lvm striping? Or more specifically, are they considered safe in this setup or will corruption slowly come about? -- Matth