On Thu, Jul 20, 2000 at 11:11:10AM -0400, Laszlo Vecsey wrote:
> I've just recently upgraded from 5.3.5 (server and clients) on GNU/Linux
> machines according to the release notes, and everything has gone
> smoothly. 
> 
> I'm still noticing occasional 'Connection timed out' errors on my diskless
> coda client, with this new version. It happens after a day or two of
> uptime on the client.
> 
> I now think it has to do with the size of the ram disk I've set aside,
> which is maxing out at 30mb:
> 
> 	/dev/ram                 30983     30983         0 100% /ram
> 
> Here is how I'm loading venus on this client:
> 
> /usr/local/sbin/venus -console /dev/console -init -vdd /ram/coda/DATA -vld
> /ram/coda/LOG -f /ram/coda/venus.cache -d 10 -h servername.domain.com &

You can set most of those things in /usr/local/etc/coda/venus.conf. Did
you remove the old container files in venus.cache? i.e. the filenames
starting with `V'. Because of the -init flag, Coda loses track of which
container files are already stored in /ram/coda/venus.cache. You might
want to add `rm -rf /ram/coda/venus.cache/[0123456789][0123456789]' to
get rid of all old container files.

Also, because everything is stored on a ramdisk, RVM is probably less
useful (in memory copy + copy on the ramdisk = twice the memory usage
and no recoverability). There is a commandline option to replace
rvm_malloc calls with ordinary malloc calls, but it hasn't been used
for a long time, so there might be some bugs still lurking there.

Jan