Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
> Dominik, what you are attempting to do is possible by way of temporarily > suspending redundancy on your cluster. And deliberately inducing split brain. > If something goes wrong on your Primary in the meantime, you will not be able > to avoid a service interruption. If you're held liable for that, don't come > whining. I can live with the risc of something happening in the meantime. My situation is as follows: I have a database's data directory on the drbd device. I saw a couple of times that - independent of drbd - database files may be corrupted for many reasons. In that situation, the secondary node wont help at all (correct me if I'm wrong, but I think I'm not). So I want to copy the files in a reasonable interval. As I can't copy the files while the database is running and I don't want to hit the primary with a dump of all data (load is high enough already), this might be a way of implementing this backup. > If you're not scared yet, read up here for inspiration: > http://blogs.linbit.com/florian/2007/10/04/checking-your-secondarys-integrity/ Thanks for that hint. > This will also tell you why a full resync is *not* necessary afterwards, > contrary to your earlier observation. :-) I didn't know "discard my data". I used "invalidate" and *that* certainly does a full sync. But that's my fault. Regards Dominik