Herbert Xu writes: This is not what the log says. The log says that the remote side told us that it is closing because it has timed out after 7200s. The 503 message apparently came from the remote server. I don't know how this can be possible if the PPP connection or the remote server is down. And the could not connect message later can only be generated when it cannot connect. That means even if you were connecting directly with a newsreader you would still fail to connect. If you have evidence to the contrary then please post them. I'll make the question clearer. The problem is not that nntpcache cannot connect and a direct connection would connect; it is that when the connection fails, nntpcache does not correctly record the attempt to retrieve the active file as `failed'. Even if nntpcache cannot connect with the news server, the update daemon finishes as if it had been successful. After a failed connect, when nntpcache connects with the news server an hour later, before the update period for the active file (4h) has expired, it does not try again to retrieve the active file. Instead, it waits until the 4h after the failed connect expires. In the meantime, nntpcache does not have a valid active file. I believe it would be more useful if nntpcache attempted to retrieve on each connection the active file until it was successful, and then started the 4h (or whatever is configured) timeout. -- Ronald Florence Maple Lawn Farm, Stonington, CT ron@mlfarm.com http://www.connix.com/~mlfarm
