home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!dtix!darwin.sura.net!spool.mu.edu!agate!ucbvax!UH01.Colorado.EDU!DWING
- From: DWING@UH01.Colorado.EDU (Dan Wing)
- Newsgroups: comp.os.vms
- Subject: BACKUP 'enhancement' in VMS V5.5-2
- Message-ID: <01GSJL128FWI006MJF@VAXF.COLORADO.EDU>
- Date: 20 Dec 92 22:35:47 GMT
- Sender: daemon@ucbvax.BERKELEY.EDU
- Organization: The Internet
- Lines: 66
-
- I'm posting the following note from DECUServe for Bob Doherty,
- <doherty@eisner.decus.org>.
-
- <<< EISNER::DUA2:[NOTES$HIVOL]VMS.NOTE;1 >>>
- -< VMS and bundled utilities >-
- ================================================================================
- Note 1962.0 VMS 5.5-2 Backup in a subprocess 1 reply
- EISNER::DOHERTY "Bob Doherty" 39 lines 15-DEC-1992 08:57
- --------------------------------------------------------------------------------
- It appears that in the process of fixing the bug in BACKUP wrt backing
- up to multivolume backup sets in a subprocess, DEC has broken BACKUP
- for some other uses in a subprocess, specifically (and unfortunately)
- the way VMSINSTAL uses BACKUP.
-
- Specifically, if you attempt to run VMSINSTAL from a subprocess to
- install, say, FORTRAN v5.9, just after the line
- Restoring Saveset A...
- the errors
- %BACKUP-F-DEVALLOC, unable to allocate device MSB0:
- -SYSTEM-F-DEVMOUNT, device is already mounted.
- are obtained, and VMSINSTAL terminates.
-
- The behavior can be duplicated by the following sequence of commands,
- assuming that the FORTRAN v5.9 tape is on MSB0:
-
- SPAWN
- %DCL-S-SPAWNED, process DOHERTY_1 spawned
- %DCL-S-ATTACHED, terminal now attached to process DOHERTY_1
- MOUNT/FOR MSB0:
- %MOUNT-I-WRITELOCK, volume is write locked
- %MOUNT-I-MOUNTED, FORT mounted on _VAX750$MSB0:
- BACKUP MSB0:FORT059.A [.VMI]
- %BACKUP-F-DEVALLOC, unable to allocate device MSB0:
- -SYSTEM-F-DEVMOUNT, device is already mounted.
-
- If the prelimnary MOUNT command is not given, and BACKUP is allowed to
- mount volume itself, the backup command works normally. Unfortunately,
- VMSINSTAL uses the separate MOUNT.
-
- This behavior does NOT occur with the BACKUP on a VMS 5.5-1 system.
-
- The version of backup which shows the incorrect behavior is X-8 with a
- link date/time of 10-SEP-1992 20:51:25.65.
-
- The version of backup which does not show the behavior is X-10 with a
- link date/time of 12-OCT-1991 04:57:31.81
-
- mumble
-
- In private Email, Bob wrote to me, in part:
-
- I have reported the problem to TSC, and they duplicated it on a 5.5-2 as well
- as a 6.0 field test system. They also said they were working on a patch, but
- it wasn't ready yet.
-
- > Where can I get X-10...
- I brought it back from our 5.5-1 system backup, right after upgrading to
- 5.5-2. From the link date (12-Oct-1991) it should be on systems bwtween
- 5.5-1 back to say 5.4-3. The version numbers don't make any sense to me,
- but they're not typos(the backup version numbers, that is).
-
- Bob
-
- --
- Posted to INFO-VAX (comp.os.vms) for Bob Doherty by Dan Wing,
- dwing@uh01.colorado.edu.
-