(Illustration by Gaich Muramatsu)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Fri, 4 Oct 2002, Jan Harkes wrote: > Cfs fr is something related to writeback caching that was never > properly finished. It pretty much always returns that error. Ah! Good to know. It did seem to flush the CMLs, but perhaps that was just lucky timing on my part, or just preamble to the operation that returns that error. > It is in fact a local/global conflict. Most likely due to a remove > operation. Yes, I did wonder how Coda would theoretically handle a conflict situation due to a removal, especially a deep one. > I guess the directory in which the conflict should appear was also > removed, so we can't show the conflict in the location it really is. > Normally the fact that there is a conflict is propagated to the parent > directory until we reach a directory that hasn't been removed yet. > However whenever there is a shell or process with that directory as > it's current working directory we cannot turn it into a symlink. Odd, because I did restart the client daemon several times as a part of testing, which involved unmounting /coda, which in turn involved killing all processes that were working in there (lsof /coda). It could very well be due to bad volume organisation on my part. When I first set all this up, I had no idea what I was doing, so I think I'm actually storing everything in what's meant to be the root volume. I have no idea what happens if a conflict were to back up that far. Can't exactly dangle-symlink /coda itself. I'm going to be building a new, dedicated fileserver soon, which will feature a more formal arrangement shared across more servers and clients. This problem had me worried I might be jumping the gun and should use NFS instead, but knowing that I can cope with minor problems like this goes a long way to reassuring me that I'll be okay. As long as I make backups. :) Thanks very much for the explanations. -----BEGIN PGP SIGNATURE----- Version: PGP 6.5.1 iQA+AwUBPZ40V/8wZ8HVeo8aEQLaCACg0IEfFIfl2wdGqJtnN4nnDoVywBEAmPUx UYvOMS4YyeU65C0TfvSSwFg= =FW3d -----END PGP SIGNATURE-----Received on 2002-10-04 22:45:31