rkl@anduin.cs.liverpool.ac.uk (10/05/90)
In article <mailnews.18010.654971633@newcastle.ac.uk>, C.R.Ritson@newcastle.ac.uk (Chris Ritson) writes: > We have got a room full of dataless workstations, a.k.a. "Diskless workstations" I assume :-) > where a student needs to be able to go to any one of them > to read mail. A sensible assumption... > The HOME directory will be NFS mounted, but hopefully not the mail > spool area. A not-so-sensible assumption ! It all depends on how your users are distributed (are you using the dreaded Yellow Pages for instance ?). The nicest solution is to soft-link /usr/mail to one directory on an NFS-mounted disk. The beauty of this is that reading mail is completely consistent across all machines that have this soft-link. The major disadvantage is if the disk (or - even worse - the machine) containing the mail becomes dismounted for some reason - then you're stuck (temp. unlinking back to a local /usr/mail is a tacky fix). > Mail will be delivered to the user's home machine. How can the user > then invoke elm and not get trouble with locking. The soft-link arrangement means then mbox.username files always get created/deleted in a consistent fashion. It will also stop 2 invocations of Elm by the same user on completely different diskless workstations. > What we could do is > give people .forward files which cause filter to save all mail in a > standard place in the user's home directory. Hmmm...I don't like the sound of this. What if your system implements disk quotaing at a low system level (i.e. it's not HP-UX :-) ) ? Won't you lose mail (or at least have to store it somewhere public - which defeats the object really) ? P.S. If you (or anyone else) are worried about unread e-mail using up precious space on one particular disk, then I could post up a mail expiry program I've written (works for BSD or System V) that expires unread e-mail after a certain period has elapsed (no, it's not a simple 'look at the timestamp of /usr/mail/username' - it actually parses individual messages and removes them if expired). Richard K. Lloyd, *** This is a MicroVAX II running VAX/VMS V5.3-1 *** Computer Science Dept., * JANET : RKL@UK.AC.LIV.CS.AND or * Liverpool University, * RKL@000010500211.FTP.MAIL * Merseyside, England, * Internet : RKL%and.cs.liv.ac.uk@cunyvm.cuny.edu * Great Britain. ****************************************************