Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
> the problem was that vgscan would detect and activate the lower level > devices instead of the "transparent virtual" block devices, > like md, drbd, whatever. Ok, so if my patch makes drbd behave just like md and nbd for lvm, it should be ok ? > > it works with lvm2, since they by design know about include and exclude > device (pattern) filters. > > maybe with drbd "internal" meta data lvm1 just works anyways, > because the lvm meta data is not found where expected on the lower > level device, and therefore not recognized. > > if you patch lvm1 to exclude certain devices in a similar fashion, > that could work, too. only, you have to get it right. > > you really do not want lvm to activate your lower level device. > > and I don't see any reason to patch lvm1, > if lvm2 is available, and more powerful. Is LVM2 available on RHES 3 (one of our targets), without patching the kernel itself ? If it's not the case, I'd rather just have to update LVM rpms and install the drbd rpm (not that I am afraid, but I'm working with very careful people :) ) Of course, on the servers where I can use LVM2, I've allready made the switch... > > > Lars Ellenberg Thanks a lot for your useful answers anyway. Marc Cousin