Here are some questions and thoughts I have on update-cluster: 1. cluster.dtd needs to be modifiable in a way to easily accomodate packages that need information other than that already present. Maybe the notes field can be used for this. For example, mosix requires a node number. /etc/exports for NFS could use options specified in a subfield of notes in cluster.xml, or the spong configuration can use a subfield of notes to determine which services to monitor. 2. update-cluster-regenerate should run an updatelist script specified as an argument to it. This way, a package can regenerate a config file using update-cluster-regenerate without bothering about side-effects. update-cluster-regenerate should run *.updatelist if 'all' is specified as an arugment. Doing that by default could be potentially dangerous. Both of these are trivial to do. 3. update-cluster-{add,regenerate} should give some help with --help. This is also trivially doable. 4. update-cluster could use debconf for input. Other packages can then call it in their postinst to generate the new config files. I'm not sure if all these are good ideas, but if they are, then I would be glad to help with the coding, as I tend to have free time nowadays. viral -- Every year is getting shorter, never seem to find the time. Plans that either come to naught or half a page of scribbled lines.
Attachment:
pgpYEWiNWusEy.pgp
Description: PGP signature