[OpenAFS] vice partition sizes, and large disks

Nathan Neulinger nneul@umr.edu
Fri, 15 Jun 2001 20:57:28 -0500

dj.byrne@jpl.nasa.gov wrote:
> > >Does it make sense to use the entire array as a single partition(~500GB)
> > >or should we break it up.
> [...]
> > btw, i really really really wish ufs logging would work with the fileserver.
> > i corrupted a couple src volumes when i tried this one.  saving the fsck
> > time on the partitions after a crash would be a big win.
> Likewise the salvage times.  Using file servers that are
> SUN Ultra-1 with disksuite RAID-5, we have 3 16GB AFS partitions.
> After a server crash (rare; usually hardware failure) we look
> at 45 minutes of downtime while the OS fscks, and THEN AFS salvages.
> Brutal for operations.  I don't have the times it would take if
> we configured as a single 48 GB partition, but assume fsck and
> salvages run in parallel, so would expect it'd be 3 times longer.

Depends on how you set up the raid - if you have them multiple
partitions on same device it seems that they do not salvage in parallel.
Although that might be a linux quirk.

We are running the FAST_RESTART and BITMAP_LATER options on our linux
servers which dramatically speed up the fileserver restart time.
Although we are still running salvager (haven't put the DontSalvage
option on yet.) I think the bitmap_later option is helping mostly.

-- Nathan

Nathan Neulinger                       EMail:  nneul@umr.edu
University of Missouri - Rolla         Phone: (573) 341-4841
CIS - Systems Programming                Fax: (573) 341-4216