[leafnode-list] Failed posting - is this right?
John Carlyle- Clarke
john.cc at europlacer.co.uk
Tue Jun 29 10:33:21 CEST 2004
I tested this, and it is repeatable. Perhaps I am covering old ground here? It is nothing to with the queuecheck script. If I send a post to alt.test during a fetchnews run, after the server which carries alt.test has been processed, then the posting will fail.
Obviously in this case, I could move my ISP's main server to the end of the list, but this still does not solve all cases. Several servers I use (e.g. gmane) carry groups not carried anywhere else.
Generally, users do not know when fetchnews is running, not currently do they get any notification of a failed posting, which may be an admin issue - I'm sure a script could be cooked up to mail the user.
I am not sure what the general solution would be to this. A configable option which prevents postings incoming during a fetchnews cycle being processed until the next cycle is one way - new posts could be put in a queue directory, which would be copied to out.going at the start of the run.
Perhaps a better approach would be to count a number of retries (per post) before failing a post. This would cope with this situation, as well as the case of a server being temporarily unreachable.
Of course, I don't know much about the workings of leafnode, so I could be miles off. Any comments?
Thanks in advance,
More information about the leafnode-list