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

Conference iosg::all-in-1_v30

Title:*OLD* ALL-IN-1 (tm) Support Conference
Notice:Closed - See Note 4331.l to move to IOSG::ALL-IN-1
Moderator:IOSG::PYE
Created:Thu Jan 30 1992
Last Modified:Tue Jan 23 1996
Last Successful Update:Fri Jun 06 1997
Number of topics:4343
Total number of notes:18308

2334.0. "*WISH* RSF HouseKeeping to tune/optimise files" by BUSHIE::SETHI (Man from Downunder) Thu Feb 25 1993 22:56

    Hi All,
    
    I have been wondering for years why the RSF (Reorganise System Files)
    housekeeping routine uses the FDL's that were originally used to create
    the files ?
    
    The problem is that we are not optimising the files at all to do that
    we should create the most up todate FLD by doing an
    ANALYZE/RMS/FDL/OUTPUT=WHATEVER.FDL. Using this FDL which is up
    todate will optimise the file.
    
    Also files like SDAF's that have records written and deleted often need
    to be optimised.  One method to reclaim buckets is to use the
    CONVERT/RECLAIM this will internally defragment the file.
    
    I would like your options and comments on this subject and correct me
    if I am wrong.  I will add this to the WISH list and hopefully it's
    included in a PFR.
    
    Regards,
    
    Sunil
T.RTitleUserPersonal
Name
DateLines
2334.1SIOG::T_REDMONDThoughts of an Idle MindFri Feb 26 1993 12:567
    Reason why?
    
    Because they do and no-one has changed the FDLs...  I logged this as a
    bug during V3.0 development, but it seems to have fallen through some
    cracks.
    
    Tony
2334.2I would NEVER use /RECLAIM here.AIMTEC::VOLLER_IGordon (T) Gopher for PresidentFri Feb 26 1993 21:1014
    Sunil,
    
    �Also files like SDAF's that have records written and deleted often need
    �to be optimised.  One method to reclaim buckets is to use the
    �CONVERT/RECLAIM this will internally defragment the file.
    
    The full $CONVERT will do this and more. Note. using $CONVERT/RECLAIM
    without first saving the input file somewhere is strongly discouraged.
    The file can easily get corrupted should some error occur during the
    $CONVERT process.
    
    Cheers,
    
    Iain.
2334.3Do you want me to SPR it ?BUSHIE::SETHIMan from DownunderSun Feb 28 1993 22:4011
    Hi Tony and Iain,
    
    Thanks for your replies.  I can see the reason why convert/reclaim is
    not used.
    
    To get this in a PFR do you want it to be SPR'd ?  Or is it already
    logged in the SPR land of wonderful things to come :-) ?
    
    Regards,
    
    Sunil
2334.4SDAF.FDLIOSG::DAVISMark DavisMon Mar 01 1993 09:5237
    
       Re the FDL file for the shared DAF (SDAF.FDL)
    
       I am not sure that creating a new FDL file everytime will gain you much 
       because if I understand you right the allocation in the FDL file will 
       equal the size before conversion which might be too large. Moreover it 
       might override sensible changes to the FDL that the customer had made 
       himself.
    
       However I am sure that there are ways that we can put better values in 
       the FDL file. An attempt at this was made in V3.0 where a size for the 
       extension is calculated based on the number of expected concurrent 
       users. The formula for extension is the number of expected users * 20 
       (or 1000 whichever is the larger). FDLs created as the result of 
       previous installations have, I think,  an extension size of 100 which 
       will lead to numerous extensions and file fragmentation. 
    
       There are some other fields that need reviewing such as
    
       Bucket size - at 4 it will hold a bucket will hold a small record 
       comfortably but over a certain number of addressees the creation of a 
       continuation record will mean that an SDAF record will span more than 
       one bucket. A higher bucket size may well be in order.
    
       Data Key compression - By default this is set to "No". This wastes a 
       certain amount of space because the key usually looks like 
       "OA$SHARn:Zaaaaaaa.aaa" whereas the key length is 65. I have heard that 
       it was originally done like this because of problems with RMS many 
       years ago. 
    
       Data fill - set at 100% . We should leave more space for the insertion 
       of new records.
    
    
       				Mark