(Insert section about processing a weekly nodelist)
This section describes how to create a nodelist modified with a nodediff, hatching out the new nodelist, then compiling the new nodelist into your BBS.
It is very important that you use three separate directories to do the nodelist processing. This is to make sure that all stages are independent of each other and if something goes wrong, you still have a working system. The three directories are:
The steps to process the nodediffs are:
These steps described in detail:
First, define the download area for the newly created nodelist. In this example, this is area 20. Users can download the nodelists and nodediffs and files to any downlinks are sent from here.
From your uplinks you usually receive NODEDIFF or NODEDIZZ files. Create a tic area for that purpose. I have keep# set to 5, this means the last 5 diff's are stored in the download directory, older ones are removed. Now you can receive nodediff files, store them for download, and send them to other nodes.
We do this with the tic magic processor. In this example I have NODELIST.007 in the /opt/mbse/tmp/nlwork directory. Note that this filename is uppercase, they are usually stored and distributed as uppercase names. As I receive the diff files as zip, the filemask on my system is nodediff.z##. This means that the file with the name nodediff.z14 in the area NODEDIZZ is a match. The command that is executed expands to mbdiff /opt/mbse/tmp/nlwork/NODELIST /var/spool/mbse/ftp/pub/fido/nodelist/nodediff.z14 -quiet if the received nodediff is nodediff.z14.
The mbdiff program applies
nodediff.z14 against NODELIST.007 in the
/opt/mbse/tmp/nlwork directory. If this is successfull, a
new NODELIST.014 is created there, a compressed
nodelist.z14 is created there and NODELIST.007 is
removed.
If this operation fails, only NODELIST.007 will stay
in that directory.
Officially, the standard archiver used in Fidonet Policy 4.07 is ARC (the predecessor to ZIP) but ARC is an ancient archiver so ZIP is normally used today. As a fallback, the mbdiff program uses the ZIP archiver to create the compressed archive. Note for Slackware sysops: if you use the SlackBuilds.org packaging system, you can find the ARC archiver in system/arc.
If creating the new nodelist fails for some reason, a missed diff or so, the whole processing stops here. The previous nodelist is still here and you can manually correct the situation. So, if you missed a diff, see that you get it and manually give the mbdiff commands as user mbse until you are up to date. Or, place the latest uncompressed nodelist in the directory /opt/mbse/tmp/nlwork.
Now that we have created the new compressed nodelist, it has to go somewhere. The file nodelist.z14 is in the directory /opt/mbse/tmp/nlwork. The example for the hatch manager is shown below. The hatch manager runs automatic with the comand mbfido tic. This setup will hatch the new nodelist in the tic area NODELIST The two screens below show the hatch and tic setup for this area.
Now that we have hatched the new nodelist and processed it the new nodelist is stored in in the download area, and maybe send some copies to downlinks. We now have to feed it to the nodelist compiler for our own system. We use a TIC magic command to do that. We unpack the nodelist in /opt/mbse/var/nodelist and set the compile semaphore so that the mbindex will compile the new nodelist. Don't be afraid that the unpacked nodelists will acumulate in the nodelist directory as mbindex will handle that and will ensure only the latest two nodelists are kept there. The mbindex program is started by mbtask.
Go Back