[OpenAFS] vice partition sizes, and large disks

[Bill Zumach]

> 
> this should improve performance when creating files.  a simple test on
> a blade 1000 running solaris 8:
> 
> localhost# mount /space
> localhost# /bin/time ./createfile 5000
> 
> real       49.3
> user        0.0
> sys         0.5
> 
> localhost# mount -o logging /space
> localhost# cd /space
> localhost# /bin/time ./createfile 5000
> 
> real        0.1
> user        0.0
> sys         0.0

That's a good start. Linux 2.2.18 is about 7,000 files, so if 3 users
are un-taring at the same time we're up to about 21,000 files. Where I'm
going with this is that you also want to check what happens when the log fills.
Irix's XFS pretty much stalls the AFS file server when it's log fills until
the transactions are committed to disk. This, BTW, is the reason for the log
size given in the instructions for making an XFS Vice partition. What I'm
suggesting is that the log needs to be sized so that it will only rarely
fill. Hm... my tests showed the AFS file server waiting using a 3.4a file
server, that is, before the threaded server. Perhaps one will not observe
this when using the threaded server. I'm sorry I can't help estimate the
required log size. The size given for XFS is purely heuristic as SGI was either
unwilling or unable to tell me how much space a given operation would take in
the transaction log and Transarc never had source for XFS so I could not
compute this myself. (Well I could have written enough tests to isolate the
value, but there was no guaranteee the next XFS patch would not change the
value).

Bill Zumach