[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference orarep::nomahs::rdb_60

Title:Oracle Rdb - Still a strategic database for DEC on Alpha AXP!
Notice:RDB_60 is archived, please use RDB_70..
Moderator:NOVA::SMITHISON
Created:Fri Mar 18 1994
Last Modified:Fri May 30 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:5118
Total number of notes:28246

5021.0. "RMU-F-INVAMBIG..../REWIND" by M5::BLITTIN () Tue Feb 11 1997 17:46

    
    Customer has two clusters. One runs RDB6.0A, the other 6.1A. Uses
    SLS to backup.  The backup command is the same on each cluster.  
    The 6.0A backup runs fine, the 6.1A fails with RMU-F-INVAMBIG, invalid
    or ambiguous qualifier [/REWIND].
    
    The command stream from the log file:
    
    RMU/BACK/SLS/OWN=SLS/LOG DB/ONLINE/CRC/NOQUIET/CHECKSUM/BLOCK=65024 -
     MUA100:DB.RBF/REWI,MUA200:/PROT
    
    The devices are mounted /for /cache, allocated.
    
    An interactive backup using the same(no SLS qualifiers) command stream
    works.
    
    Appears that the 6.1A backup process thinks the tape device is a disk?
    
    Is there something that I can look at to confirm this?  Is this a 
    restriction somewhere?
    
T.RTitleUserPersonal
Name
DateLines
5021.1BOUVS::OAKEYI'll take Clueless for $500, AlexTue Feb 11 1997 17:5111
~~                       <<< Note 5021.0 by M5::BLITTIN >>>
~~                         -< RMU-F-INVAMBIG..../REWIND >-

Bob,

What version of SLS?

Also, just to make sure I understand... you're saying that if you remove 
the /SLS qualifier from the command below, it works with both versions of 
Rdb?

5021.2DUCATI::LASTOVICAIs it possible to be totally partial?Tue Feb 11 1997 18:0016
As you suggest, it appears as though the RMU code believes that
a tape isn't being used so the /REWIND qualifier isn't valid.

Sure looks to me as though RMU believes that there isn't a tape
involved.  you might want to examine a SHOW DEVICE/FULL on 
MUA100 & MUA200.

>/PROT

seems odd to specify the protection switch without a protection mask.
though I doubt that this has anything to do with the problem.


>    Is there something that I can look at to confirm this?  Is this a 

	are you able to reproduce this?
5021.3HOTRDB::PMEADPaul, [email protected], 719-577-8032Tue Feb 11 1997 19:171
    Are you sure they have the proper DCLTABLES.EXE visible on all nodes?
5021.4Some clarification...M5::BLITTINWed Feb 12 1997 10:1234
    
    1.  Will get SLS version...
    
    2.  Exact command stream as it appears in log:
    
    $ run/NODEBUG sls$system:sls$dbsysbak.exe
    %MOUNT-I-MOUNTED, RM0130 mounted on _HJRM12$MUA100:
    %MOUNT-I-MOUNTED, RM0130 mounted on _HJRM12$MUA100:
    %SLS-I-STARTING, starting volume RM0130 at position 0 on drive
    _HJRM12$MUA100: at 23-JAN-1997 00:48:29.12
    $ V = 1
    $ SET NOON
    $ ASSIGN _FTA7: SYS$COMMAND
    $ ASSIGN _MBA1773: SYS$ERROR
    $ SET NOVERIFY
    $ RMU/SHOW VERSION
    Executing RMU for DEC Rdb V6.1-1
    $ RUN/NODEBUG SLS$SYSTEM:SLS$LOAD_FINALSTS.EXE
    $ -
    RMU/BACKUP/SLS/OWNER=SLS/LOG -
     -
    DSA3:[PRDBD.SUBRPT01.ROOT]DBS_SUBRPT.RDB;/ONLINE/CRC/NOQUIET_POINT -
    /CHECKSUM_VER/BLOCK=65024 -
    _HJRM12$MUA100:SUBRPT_23JAN.RBF/REWI,_HJRM12$MUA200: -
     /PROT=(S:RW,O:RW,G:R,W:R)
    %RMU-F-INVAMBIG, invalid or ambiguous option "/REWIND"
    %SLS-I-FINISHED, finished volume RM0130 on drive _HJRM12$MUA100: at
    23-JAN-1997 00:48:33.61
    
    3.  I'll check on the DCLTABLES, but would expect it to be ok since
        they can run this interactively...and there could also be two
        different copies in diff places...I'll check.
    
    Thanks...
5021.5Problem found...thanksM5::BLITTINWed Feb 12 1997 11:209
    
    Customer said they found problem...
    
    Stated the command procedure that builds the parameter file for SLS 
    defines a symbol DRIVE_TYPE :== "HJRM12$MUA100:,hJRM12$MUA200:", 
    but left off the last '"' (it is missing), but I don't understand
    how this caused the problem.
    
    Their happy....
5021.6NOVA::DICKSONWed Feb 12 1997 11:406
    Part of the algorithm by which RMU decides if a backup file is going to
    a tape or not is very sensitive to the syntax of the filespec.  The VMS
    system service "SYS$PARSE" seems to think that a device name that does
    not end with a colon is really a filename.
    
    And you cant specify /REWIND with a filename.