NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/os/vms/18224

Text File | 1992-11-19 | 1.6 KB | 35 lines

Newsgroups: comp.os.vms
Path: sparky!uunet!ornl!rsg1.er.usgs.gov!darwin.sura.net!aplcen.apl.jhu.edu!procon!rubin
From: [e-mail] (Don Rubin)
Subject: Re: Wierd BACKUP change in VMS 5.5-2
Message-ID: <[e-mail]>
Lines: 22
Sender: [e-mail] (USENET News System)
Organization: Johns Hopkins Univ APL
References: <[e-mail]>,<[e-mail]> <[e-mail]>
Date: Thu, 19 Nov 92 17:53:23 GMT
Lines: 22



!If memory serves, you've *NEVER* been able to mount a tape that was allocated
!to the parent of the current process.  VMS v4.0 made some changes in the way
!tape mounting was done that allowed the tape to be shared among processes in a
!job.  If I recall correctly, that version of VMS had a bug that caused the
!MOUNT to also allocate the tape (or maybe it changed the ownership of a tape
!allocated by the subprocess to ownership by the parent; it's been a long time
!since 4.0 and my memory's a bit fuzzy).  Thus if one mounted a tape from a
!subprocess, then dismounted it, a subsequent attempt to mount it would fail.

I was the original poster. The problem occured in moving from VMS 5.5-1 to
5.5-2. Despite the above comments, on our system, the sharing of a mounted
tape did work. That is, if in the main process I mount a tape (from code
using sys$mount) then SPAWN BACKUP in a sub-process (using smg$create_subprocess),
it would work fine. This stopped working in vms 5.5-2 (at our site). DEC
support has referred this to engineering.

--
Don Rubin
Johns Hopkins Univ APL
[e-mail]