On Thu, Apr 27, 2000 at 12:33:22PM +0530, Madhav wrote:
> Hi all,
> I am getting this error when I gave "ls" under /coda  in a coda client when
> the scm is down.
> Even after the scm comes alive and I logon(clog) also it gives the same
> message. But before the scm is down I didnot logon(clog) into client. Am I
> missing something. Can anyone please help me.
> 
> Thanks in advance,
> Madhav
> 
> [root_at_server2 /coda]# ls
> ls: qmail: Connection timed out
> ls: ftp: Connection timed out
> ls: srv1: Connection timed out

The attributes of those files are not in the client's cache. ls (-l)
wants to get the size/mode/uid of the files, and venus cannot provide
them, so ETIMEDOUT is returned.

After you reconnect to the network (or bring up the scm) the userspace
code isn't told about this fact, and relies on regular server probes to
detect that the `down' server has become reachable again. That is one
cause for getting ETIMEDOUT after reconnecting the client.

You can force a serverprobe by running 'cfs cs'. The other cause for
continued disconnection is when there is a conflict in the volume. For
some reason I haven't figured out yet, the client goes to disconnected
mode (instead of write-disconnected) when a conflict is detected. The
volume state and whether there are possible conflicts can be seen using
`cfs lv /coda'.

Jan