Ron, Thanks for your second reply to this subject.
Cost is certainly an issue between optical disks and hard disks.
Hard disk prices are dropping, and the reliability is there.
I am still looking at optical disks, which may drop
as well. I want to keep my options open at this point.
Your first reply last week, I recall, talked a lot about problems
I would have with thrashing i.e. having the wrong optical disk
mounted which could cause network problems while it got mounted.
The Andataco cache disk in front of the optical disk may
eliminate much of this contention.
You included a note from Lyle Seaman with comments about the size
of a partition.
The 2 GB/partition limit is fundamental, I think. It's an artifact of
the interface: the section 2 routines all take "int" offsets and return
signed values. We'd have to really overhaul the file server to
eliminate that.
This is a very tentative comment. The part about
the "section 2 routines"
is not clear. I think he is talking about the
system calls like "lseek". If so these limit the size of a FILE
to 2GB, but not necessarily the size of a partition. There may be
a limit to the size of a volume, which I could live with at 2GB.
Its the size of the partition that I am interested in.
Most UNIX systems these days can have file systems greater
then 2GB. Is there really some other problem with AFS?
Since most references to partitions are to blocks, and not to
bytes, a long int reference could be covering 512 or 1K byte
blocks, or 512GB to 1024GB partitions.
I would like to get him, or someone else at Transarc to look
into this. (We do not have the source.) I have not found any
reference in the manuals to the size of a partition.
Douglas E. Engert
Systems Programming
Argonne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439
(708) 252-5444
Internet: [EMAIL PROTECTED]
