> > > Regarding faulting the RAID to take one of the drives as a snapshot: > be > > careful that reconstruction of the new drive puts a non-standard > stress on > > the drive that's still in the system (higher the bigger is the > drive), and > > it may increase the risk of it dying during the process, taking down > the > > whole system. > > I've never heard of this concern before. It seems like an unlikely > scenario, since the "stress" is simply reading the remaining drive.
It is unlikely, but the stress is not minimal: it's reading several TB of data at full speed for hours, non stop. If the drive is close to failure that may tip it over the edge right in the moment in which the system is vulnerable. But in any case, if that were to happen, the failed drive would be > just one of two copies of the data (the other being the one I had just > removed), and the running system would still be healthy as well. > True, but that drive is not installed in the system, so there is downtime associated (it may be relevant or not). And worse, you have now only a single copy of the data, with again a chance of failure while building the mirror (and that chance is not negligible if the disk comes from the same hardware batch). Of course, you can do things to minimize those risks, but my point was that one should be aware of those extra risks and balance if they're important or not, compared with other ways of backup that may be safer but more cumbersome. Regards, Guillermo
_______________________________________________ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List: https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki: https://github.com/backuppc/backuppc/wiki Project: https://backuppc.github.io/backuppc/
