nlehrer@ibb.gov (neil lehrer) writes
i understand that having more than 1 raid grp in a filesystem makes the filer more fault tolerant of a full raid grp failure.
If by "a full raid grp failure" you mean the failure of two discs in the RAID group, then: no, having more than one RAID group in a volume doesn't make it more tolerant of such a failure. But it does make such a failure less likely (because, if two discs fail, they are not necessarily in the same RAID group).
That's not the only reason to limit the size of RAID groups, of course. There's the diminishing returns from being able to do full stripe writes when the stripe is too large.
so lets say there are two raid grps in the filesystem. if you lose a raid grp [2 drives fail] how do you restore it? i can understand restoring the whole filesystem or a qtree or a dir, but if files and dirs can go across more than 1 raid grp in a filesystem what is the advantage of short restore of the broken raid grp? how do you put back just that raid group?
i know you can replace the bad drives and then recreate the group but how do you restore the data for just that raid grp?
You can't dump or restore by RAID group at all. If you lose a RAID group, you lose the volume, and will need to restore it all from backups, or mirrors, or whatever.
Chris Thompson University of Cambridge Computing Service, Email: cet1@ucs.cam.ac.uk New Museums Site, Cambridge CB2 3QG, Phone: +44 1223 334715 United Kingdom.