Conference iosg::all-in-1_v30

    Hoi Martin,
    
    How is live in the far south of Holland ?
    
    I ran FCR with set watch. Do the same and compare it with 
    FILE_CABINET_REPORT_SM.LOG1992111220000000 on UTRYIT.
    
    See you,
    
    Jan
    
    CSC Holland Central
                                      
    

Hi jan,

I already did run FCR with the set watch utility.

The customers has four mail areas OA$DAF_A.DAT, OA$DAF_B.DAT, OA$DAF_C.DAT and
OA$DAF_E.DAT.

First I had the situation that after opening correctly OA$DAF_C.DAT I've got the
error mentioned in .0

Last weekend the system manager rebooted the ALL-IN-1 cluster via autogen 
because he needed to adjust some systemparameters.

Yesterday I ran FCR again and the message did appear after opening correctly 
OA$DAF_B.DAT.


Maybe, there's something wrong with system parameter values


Greeting from the rainy but most of the times sunny south.

Martin

    The error is coming from a sub-routine that attempts to open all of the
    known SDAFs on a system.
    
    Is there a duff record in the SDAFs master file?  For example, is a
    record pointing to an SDAF that doesn't exist or is in the wrong
    location?
    
    Tony
    
    Code extract:
    
    1600    !   ALL-IN-1 V2.3 allows a system to have up to 5 SDAFs.  To
            !   find out which ones are available on this target system we have to
            !   open the SDAF master file and read down through it to see what's
            !   around.
    
            !   Once we find something we can open it.
    
            until end_of_sdaf_master%
            get #11%
            daf_file_spec$ = trm$(daf_location$) + trm$(daf_file$)
    
    1601    select daf_area$
    
    1602    case "OA$SHARA"
               open daf_file_spec$ for input as file 15%             &
                 ,organization indexed variable                     &
                 ,allow modify                                      &
                 ,access read                                       &
                 ,map daf
                 daf_a% = 1%
            case "OA$SHARB"

Hi tony,

>    Is there a duff record in the SDAFs master file?  For example, is a
>    record pointing to an SDAF that doesn't exist or is in the wrong
>    location?
    
No, there's isn't a duff record in the SDAFs master file.

But there's an error in FCR !!!!!

I can even reproduce the error.

I've a VAXstation with VMS 5.4-3 and ALL-IN-1 V2.4 (patch K603) and FCR.

I did the following tests:

1. Ran FCR with OA$SHARE active --> no problems

2. Created OA$SHARA, ran FCR with OA$SHARA and OA$SHARE active --> no problems

3. Created OA$SHARB, ran FCR with OA$SHARA, OA$SHARB and OA$SHARE active -> no
   problems

4. Created OA$SHARC, ran FCR with OA$SHARA, OA$SHARB, OA$SHARC and OA$SHARE 
   active --> Error

5. Delete OA$SHAREC, ran FCR with OA$SHARA, OA$SHARB, OA$SHARE active -> no 
   problems

6. Created OA$SHARD, ran FCR with OA$SHARA, OA$SHARB, OA$SHARD and OA$SHARE 
   active --> Error

7. Delete OA$SHARED, ran FCR with OA$SHARA, OA$SHARB, OA$SHARE active -> no 
   problems


So running FCR with more than 3 mailareas open results in an error message.


Martin Smeets

    I'll check this on my system when I get home, but FCR V4 is out of
    support (not supported) now that FCR V5.0 has shipped for ALL-IN-1
    V3.0.  Actually, FCR V4 was never supported as it was always provided
    as a Solution Tool from ASSETS.  FCR V5.0, on the other hand, is a PASS
    and comes with support.
    
    Tony

    I have now run FCR V5.0 on ALL-IN-1 V3.0 with 5 SDAFS (A through E)
    open and active. No problems occurred.  
    
    I know that this isn't very helpful (the next version always fixes a
    problem), but can you tell me what minor version number is reported by
    FCR V4.1?  There may be alater version of V4.1 available where the
    problem was fixed.  I sometimes took fixes across into V4.1 after
    making them in V5.0, so maybe I can give you a new version...
    
    Tony

Tony,

It's FCR V4.1-003.

Regards,

Martin

    The latest code I have for FCR V4.1 is V4.1-006 dated 2 March 1992.
    
    I can produce a new version of the program if you want, but it will
    have to wait until I get home (sometime next week).  Is this for a
    customer?  If so, will they pay ;-)
    
    Tony

Hi Tony,

Yes it is a customer, but I don't think they will pay !

Martin

    Well, then they'll have to wait until I get the chance to do some
    unfunded work.  I should manage to do this in late 1994.
    
    Tony

Hi Tony,

> Well, then they'll have to wait until I get the chance to do some
> unfunded work.  I should manage to do this in late 1994.

Maybe you don't have to do any unfunded work. Could the error been a result of
low pagefile or badly fragmented pagefiles ?

Martin

    Maybe. I know that I tested multiple SDAFs when the program was
    written, so maybe it is due to other system influences... like drink
    (too much on the part of the system manager), or some other reason.
    
    Tony

Finally, I've found the error. The problem was caused by a low GBLPAGFIL value.

So raising the value did the job.

Martin