home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.protocols.nfs:2231 comp.mail.elm:2310 comp.sys.sun.admin:6148
- Path: sparky!uunet!pipex!warwick!uknet!liv!qq11
- From: QQ11@LIVERPOOL.AC.UK (Alan Thew)
- Newsgroups: comp.protocols.nfs,comp.mail.elm,comp.sys.sun.admin
- Subject: Re: Frequency of NFS lock problems w/ NFS mounted mail spool
- Message-ID: <92248.145904QQ11@LIVERPOOL.AC.UK>
- Date: 4 Sep 92 13:59:02 GMT
- References: <1992Aug27.185938.21422@leland.Stanford.EDU>
- <vanepp.715392389@sfu.ca>
- Organization: University of Liverpool
- Lines: 48
-
- In article <vanepp.715392389@sfu.ca>, vanepp@fraser.sfu.ca (Peter Van Epp) says:
- ....
- >>For several years, I have run small networks with /var/spool/mail
- >>NFS-mounted from a single server to a bunch of workstations. I have
- >>yet to have a problem doing this. Most of these machines have been
- >>Sun's running a recent version of SunOS.
- >
- >>However, as I understand it, due to various bugs in various vendors'
- >>rpc.lockds, this is running the risk of getting one's mailbox
- >>thoroughly hosed if more than one process tries to write the mailbox
- >>at once.
- ....
- >>I'd be curious to know how many folks have actually had this kind of
- >>problem -- if it turns out to be very few, NFS seems like it would be
- >>worth the convenience. Specific experience about the reliability of
- >>the SunOS and Ultrix lockd's would also be interesting.
- >
- >>Also, where do the Elm lockfiles fit into all this? Would they
- >>prevent this problem (if everyone were to use Elm, that is), or do
- >>they merely keep multiple copies of Elm from manipulating the same
- >>mailbox simultaneously.
- ....
- >
- >As I recall (it has been about a year since I was through this) Elm uses
- >flock()not lockf(), so NFS locking isn't a problem (nor possible for that r
- >matter since
- >flock() won't work on an NFS file), what does the locking is the .lock file t
- >that
- >Sun Sendmail creates in the mail spool directory. We have some 10,000 or so
- >active users from 6 servers all accessing a mail spool NFS mounted on an
- >Auspex
- >without any problem. Again as memory serves, Elm is prepared to use the .lock
- >file if asked to.
- >
- [ sorry for including so much quoted stuff but it all seemed relevant, I have
- also reduced the number of newsgroups ]
-
- According to elm's config.h, the default is to use flock() and also .lock
- files. So if I cannot trust rpc.lockd, then my MTA must use .lock files as
- well and have a common lock directory with elm. The comments in config.h
- are not clear whether this is /usr/spool/locks or /usr/spool/uucp...
- (the defaults are based on Elm's Configure defaults).
-
- Thanks for any futher input.
-
- Alan Thew : Computing Services Department, The University of Liverpool
- Internet/Bitnet : qq11@liverpool.ac.uk UUCP : ....!mcsun!uknet!liv!qq11
- Voice : +44 51 794 3735 Fax : +44 51 794 3759
-