-*- Text -*- * Creating a CCLAN archive node 0) Pick a place Requirements: - some piece of filesystem that other people can read - sh, sort, uniq, awk, rsync and perl - if you want to run a node that can be uploaded to, the rsync server We'll call the archive directory $CCLAN. If you plan to allow uploads, it needs to be accessible using the rsync server 1) Untar cclan-skeleton-x.y.tar.gz into $CCLAN 2) Edit the definition of $CCLAN in $CCLAN/bin/update 3) Create the file $CCLAN/etc/peers with rsync:// URLs to the other cCLan node(s) that you have arranged to synchronise against. One URL per line. e.g. $ cat etc/peers rsync://ftp.linux.org.uk/ftp/pub/lisp/test/ rsync://glodfish.telent.net/cclan/experimental/ $ (P.S. This example probably doesn't work. The second line especially so) Different node owners may have different policies regarding mirrors. There should be a doc/PEER-POLICY file in each node which explains what the policy is for that node and who to contact - please do not attempt to peer with a node without its blessing. 4) Edit your own doc/PEER-POLICY file which expresses your policy on peers. Optionally add a doc/README file with general information about your node to be presented to archive users, and add a symlink $CCLAN/README -> $CCLAN/doc/README 5) Run $CCLAN/bin/update to populate the archive. Arrange for it to be run regularly henceforth (e.g. using a crontab entry) 6) If you want to upload files into your new node and have them propagate throughout the network, you need to find node owners who will add you to their etc/peers. Try asking on cclan-list@lists.sf.net Note that these scripts only provide for the peer to peer propagation aspect of uploads, not for the initial posting of content onto your node. How you do that - shell tools, ftp wth anonymous put, or whatever else - is a function of your local policy. Get the files into $CCLAN and we'll take care of them from there. * Node Maintenance The update script does not delete files. Ever. If you enable the -m option to fix-symlinks it will move them to old/ from where you can delete archives. Don't delete the corresponding .meta files until they're _really_ old (if ever) or you may find that old archives come back from your peer nodes after you thought they were gone.