On 11/12/2011 07:19, Jason Hellenthal wrote: > If end-user is upgrading a package they should be prepared to take > any neccesary action to start the services again after final actions > are complete. Desperate services could have a periodic script that > could handle the checks for these services and I strongly advise > against taking any action whatsoever on starting or restarting or > reloading.
Like this? http://www.infracaninophile.co.uk/articles/hotchpotch.html#status-rc Not that I entirely disagree with having a mechanism available to restart daemons on update, but yes, it should definitely be optional and configurable and the docco should explain very clearly that *it will not work properly in all cases*. There are too many corner cases where you've got to worry about the order in which services are restarted, plus things like needing to restart a completely different service when certain ports get upgraded. Consider what needs to happen if you update a commonly used shlib to fix an important problem but without there being an ABI bump... Cheers, Matthew -- Dr Matthew J Seaman MA, D.Phil. 7 Priory Courtyard Flat 3 PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate JID: matt...@infracaninophile.co.uk Kent, CT11 9PW
signature.asc
Description: OpenPGP digital signature
