Hi Ivan!
>
> let you take it cautiously.
Well, I do - for testing purposes I only moved my own home...

>
> > Things seemed quite fine until I started testing the whole thing with
> > files > cachesize.
Actually it seems as if it had all been my own fault: The partition
containing the coda-cache had run out of space! Still venus shouldn't
crash in those situations...

> Coda is not supposed to be used with file sizes bigger than cache size.
> Whole file caching is fundamental for Coda, and you have to have space
> for at least all simultaneously open files.
But files > cachesize _will_ exist and users want to be "unlimited"!

> > There seems to be no problem if I simply copy a single
> > file (larger than the cachesize) to a coda-volume, but when I checked out
>
> It may work, but usually it does not make much sence as you never can
> access that file from that client.
Wouldn't it be possible to implement full access as long as you are
connected?

> > CML-reintegration-problems (btw. what does CML mean?), that did not fully
>
> Cache Modification Log (?)
Thanks!

> > Furthermore venus seemed to be crashed (`cfs cs` and ctokens hung) and a
> > restart of venus was only possible after purging the DATA & LOG -files.
>
> I suppose you might have hit some bug in venus by asking it for impossible
> operations, or just by the CVS program which stresses the filesystem in
> a special way.
> (CVS essentially is not compatible with Coda - it relies on exclusive
> access/locking which in general does not exist on Coda.
> I do not say it cannot be used, but I suppose its usefulness is very limited.)
But that is only of interest, if the repository itself had been on a
coda-volume, which is not the case - I just checked out a working copy!

Regards,
Michael