Hello all,

I've setup coda 6.0.2-1 a few months back as a non-replicated server on
a desktop (with no other volumes aside from the root volume) accessed by
two laptop clients over a lan. I plan to observe large file transfers
using coda between the clients and the server once i change this setup
into a wireless one (something to do with a study i'm conducting..but
that's another story). Anyway, since I've been lurking around this
mailing list for some time, I'd like to share my own thoughts on my
experiences with coda so far since I spent some time tinkering and
experimenting its stability.

It seems like I have had a relatively smooth time using coda compared to
others according to the mail archives, but then again it might be
because I have very straightforward setup. In any case, I'm quite happy
with what I have...aside from academic reasons, my boardmates and I use
coda as an alternative to NFS for casual file sharing in our small LAN.
The only problems that I've been getting are the "local inconsistent
object" errors by venus which tend to occur after I intentionally
perform heavy disk operations. I tend to be able to recreate such errors
by untarring and tarring archives with around 85 files, deleting, moving
and basically fiddling around with them. It may or may not mean
anything, but untarring and tarring archives with numerous symlinks in
them seemed more likely to cause local inconsistent object errors than
archives that didn't. I'll leave the experts to make the call...but
chances are that this is just a coincidence.

Btw, just in case the info might come in handy...my RVM is stored as a
315mb file (alas...not a raw partition), i never store more than 400mb
in the coda partition, and the number of files in the partition don't
usually exceed 300. Venus cache sizes are all 100mb. The clients and
servers all use a coda-realm patched 2.4.20 linux kernel to accomodate
coda 6.0.X.

Some questions...

-testserver.coda.cs.cmu.edu has always been recognized as a realm under
my /coda. Just recently however, i've constantly been getting numerous
"23:13:25 strauss.coda.cs.cmu.edu nak'ed" messages. I read from the
archives that this has to something to do with the connection to
strauss.coda.cs.cmu.edu, but i was wondering if somebody could explain
what this really means and what the implications are. Also..why is it
happening now? :) Oh...I'd like to apologize in advance if this has been
discussed before.


-I'm interested as to how modifying the cache object priorities in the
hoard database (like when the hoard command is used) exactly affect
overall caching. What is the "default" priority value for an object
then? I'm assuming that there exists a default priority value of lets
say "500", and modifying a file priority would tell venus something like
"cache this file at all costs, and only give way to files assigned
higher priority values". Of course I'm willing to bet there are more
factors involved in the caching internals. I manage to browse the source
from time to time, but I'd appreciate it very much if someone could give
me a brief overview on how this exactly works or point me to some
documentation.

Ciao for now, and more power.


=============================
The statement below is true.
The statement above is false.
=============================