Re: Design of NLM

New Message Reply About this list Date view Thread view Subject view Author view Attachment view

From: Eric Werme USG (werme@zk3.dec.com)
Date: 01/09/03-12:56:42 PM Z


From: Eric Werme USG <werme@zk3.dec.com>
Message-Id: <200301091856.h09Iugo0001612064@anw.zk3.dec.com>
Subject: Re: Design of NLM 
Date: Thu, 09 Jan 2003 13:56:42 -0500

   One some UNIX Flavors like Tru64, IRIX etc there seems to be no 
   lockd or statd and I suppose that tells that they do not have 
   the Lock Manager or Status Monitor Protocol implemented.

Not at all.  As Guy Harris mentioned, "rpcinfo -p" is authoritative,
though I prefer "rpcinfo -u" or "rpcinfo -t" because that actually
contacts the service by sending a NULL request.

In Tru64's case, we have supported locking since day one:

  $ ps ax | grep rpc
   485 ??       I        0:00.03 /usr/sbin/rpc.statd
   487 ??       I        0:00.07 /usr/sbin/rpc.lockd

Our sysadmin tools allow _not_ running them, so keep in mind that
rpcinfo will reports only that the service is currently available.
It has no way to check whether support has been implemented but is
not in use on a particular system.

Irix also supports locking.  In general, any credible UNIX NFS
implementation has locking, and any NFS toaster sold in the Unix
market has locking.

 While 
   on Solaris I do see these. In case these protocols are not 
   implemented, then how is locking achieved and what type of 
   locking probably they would follow.

There are some classic (and mostly deprecated) locking forms, but
they generally lock an entire file, don't work well over reboots,
and in general cause more problems than record locking does.

This is off-topic for the V4 mail list.  If you have other questions about
NFS in Tru64 Unix, please contact HP support or me directly.

	-Ric Werme


New Message Reply About this list Date view Thread view Subject view Author view Attachment view

This archive was generated by hypermail 2.1.2 : 03/04/05-01:50:46 AM Z CST