Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Tue, Mar 09, 2010 at 10:59:54AM +0100, Olivier LAMBERT wrote: > I've got your point. I know my first message is long and maybe > complicated. It seems you "jump" some lines. It's not "nested" LVM as > described in the doc (because I RTFM before try to ask help). > > My problem is not located on DRBD hosts. I've run some test, and I can > really simplify the example. Please, read this, I think it's clear now > : > Let's say I export a simple physical volume with DRBD. So, NO LVM in > DRBD, just replicate "/dev/sdb" for example, and export it > (/dev/drbd0) on each DRBD host with iSCSI. > > So, on ALL my clients (connected iSCSI), I saw with "fdisk -l" : > Disk /dev/sdb: 2998.8 GB, 2998869356544 bytes > > So, my CLIENT1 connected to DRBD1, decide to create a volume group and > a logical volume. > Despite all my efforts (vgscan, pvscan, lvscan), CLIENT2 connected to > replicated DRBD2, cannot see anything ! No VG, no LV ! Nothing ! > > But, if I decide to do that on DRBD2 : > drbdadm disconnect ressource && drbdadm connect ressource > > Just after THIS, I can see all the stuff I've created on CLIENT1. > > So, it's NOT a question of LVM on DRBD, because in this case, I don't > use it anymore. I can undestand if LVM data's on client are not > replicated on each DRBD hosts, but WHY it works if I > disconnect/reconnect these two ? It doesn't have any sense to me.... Does the keyword "cache coherence" help? -- : Lars Ellenberg : LINBIT | Your Way to High Availability : DRBD/HA support and consulting http://www.linbit.com DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. __ please don't Cc me, but send to list -- I'm subscribed