Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
> I don't see why a backup on the secondary node would not work > as long as the > volume you're snapshotting is underneath DRBD. In fact, I've > done just that. > The data is there after all. Only problem is, that it's in an > unknown state. > There may be writes in progress. But if your applications can > reliably cope > with this (like e.g. a good database should), it's certainly > an option. Yeah, I figured it's no worse than recovering after a power failure. When the snapshot gets mounted, it will do the normal filesystem recovery. We contemplated using a "real" shared filesystem in dual-primary mode, but I assume that the shared filesystems don't perform as well... just a guess. lew