-- Back Home
-- calendar
-- Jobs to do --
-- Previous week
-- Next week --
|
-- Back Home
-- calendar
-- Jobs to do -- |
1999-03-15, Monday
Together with Peter I’ve been checking the Italian system and our programs. We don’t find a way to explain it. The error occurred in IVC022R, where the OK file gets copied to the KI file. At the time of copying the OK record could not be found, we know that because there is no KI record so no conditions were applied. Problem is that we see the record in KO/OK file and it should have been found.
There is NO PROBLEM in Italy (except the Italians of course). The invoice(s) they said were still wrong dated from the 9th but we did sent the patch on the 10th. After installing the patch they did correct the OK/KO file and ran the invoicing. The invoicing wil only recalculate invoices that aren’t calculated yet so the erroneous invoices were not corrected. New invoices should be correct. Upon informing them of our suspicion in that direction they confirm this, as if they knew all along.
“We agree with you. Actually, the invoicing run on Friday evening results
with no-errors. We will maintain our daily controls on order-invoice
differences for another weeks at least.
We will inform you in case of wrong invoices.”
So they knew the programs were correct and didn’t inform us that the newly reported problem was solved. This did cost us a day!
Sigh! Now that the Italians are handled the Germans begin to call for help.
They are wondering (=they thing there is an error) about the handling the GE file and postings to OFI.
Die Datei GE wird aus der Datei HI mit Teildatei HIOM3 (postet to GL <> Y) gefüllt.
We checked it here, in the programs. What they imply is not true. The records are deleted from the HxOM3 member only after both the invoicing (ivc002r) and the GL-posting (ivc025r) has run. The GL posting program checks the OHSTS field. Only if it is ‘CLS’ (order is closed) the records will be deleted (by a call to OCR710R). An order is closed by the invoicing program. The invoicing program checks HIPGLI (posted to the GL interface) field. It only deletes records when that field contains $yes (translatable). $yes is put in there by the GL-posting program. HIPGLI is also set to $yes in the invoice-calculation program (inv022r) for invoices with urgency code 9. That is for non-goods invoices (credit notes and returns). This was requested in PRS 9718.
Die Teildatei HIOM3 wird bereinigt, sobald aus der Vor-Faktur eine tatsächliche Faktur gemacht wird.
Da wir über den ganzen Tag Einzel-Fakturen anstossen (meist Gutschriften/Rückholaufträge) fehlen diese Umsätze in der Tagesabschlußverarbeitung und werden nicht nach OFI (F58010) gemeldet.
Lösung?: für das Füllen der GE-Datei muß nicht nur die Teildatei HIOM3 auf das posted-Kennzeichen geprüftwerden, sondern auch die HI-Teildatei.
Bitte prüft dieses Problem und nennt mir schnell eine Lösung.
A possible problem is that the GL-posting program uses 'CLS' as a hard coded value in stead of a local-translatable value that gets retrieved from the CD file. If that proves to be a problem we can change it.
Adding some Windows tips to the windows file is all I came to this evening.
1999-03-16, Tuesday
Peter is on a SAP course.
I’ve been catching up on mail and view. There was a mail and phone from Germany. They have a problem with posting GL records (see yesterday). I searched and found a PRS that is related to that ‘problem’. This is not really a problem with the programs but rather a problem with the specifications.
I ‘invented’ a new recipe for lasagna to day. Using haddock as main protein source and a paprika based sauce. We will have to wait till somewhere next week to get to taste it but I have good hopes for it. If the taste is right I’ll post it here so stay tuned. If this site suddenly stops next week you’ll know that haddock with paprika in lasagna is a deadly combination!.
1999-03-17, Wednesday
I have spent a lot of time today on administration and bookkeeping. The PRS and problem documentation is up to date again.
The search on TH goes on.
I just got a call from Francois Vestjens, from Germany, about the posting GL problem. He is convinced that this must be solved ASAP. The PRS was agreed on during a meeting with John Knoops, Fred Peeters, Jan Geurts and Peter Hilberts after the integration tests because credit notes without goods flow entered OFI twice, once via the F58010 and later via F5806x. Now Francois wants to reverse the situation because the data should reach OFI immediately upon invoicing.
I just added a few remark to the UNIX files.
1999-03-18, Thursday
First thing in the morning I tried to install Win98 on Yaku. Previously it gave the error that a compression program could not run in this release (win98) and stopped death. Now Yaku has a 2.5GB primary disk and a 2.1GB secondary disk, both in IDE chanel-1. There was no compression used. The only possible problem was that this primary disk used Bootmanager from IBM to allow booting in DOS (first 60MB) or Windows95 (from 61 to 1500MB) or OS/2 Warp4 (rest of the disk) with the boot manager residing in a primary 2MB active partition at the end of the disk.
I removed the DOS partition and extended the Win95 partition to the start of the disk (using PartitionMagic). This doesn’t solve anything. Same error.
Next I removed everything from the primary disk. Now the Win98 boot floppy doesn’t find a CD drive and obviously can’t continue. I am left on a a:> prompt with ver responding with “Windows 98 [Version 4.10.1998]”. The active partition, before I low-level formatted, was the boot manager partition. This was a FAT partition but located at the back of the disk, further than 2GB so maybe the FAT16/BIOS combination limit of 2GB disks caused a problem that the Win98 install program stripped over thinking it had hit a dickcompression utility.
That my CD drive is not recognized is strange. This is a Mitsumi CR-2801TE drive that is recognized by the BIOS (mode-3). The installation tries to load OAKCDROM.SYS with OEMCD001 as parameter but this fails, saying “No drives found”. Later other drivers are tried (btdosm.sys; flashhpt.sys; btcdrom.sys) none of which give any messages. Then ASPI drivers from Adaptec (including aspicd.sys) are tried, all giving version and copyright messages but no success or failure information. At a certain point in the config.sys an if statement is encountered (if “cd”==”nocd” goto quit) and then the autoexec is started. Here MSCDEX.EXE is tried but this cancels with “no valid CDROM device drivers selected”.
As last resource I’ll put a standard simple CD drive in and try again. Tomorrow.
Ronny wants to test the release to hold process in TS2000. For this he checks in the DTAQ file to find out what OLI is doing/has done to the order. Testing he never gets to a filled DTAQ file because this doesn’t keep data persistent. When a job is put on the data queue a record is added and when the job finished that record is removed again. It happens all to fast for testing.
After countless interruptions some tests could be performed. To do the tests I looked up what gets filled in in the DTAQ file, then I stopped the background jobs to prevent true processing. Next I manually added the same things in the file and let Ronny execute his program. After the test I removed the record again from the DTAQ file.
This was repeated for each option.
No problems cropped up.
I keep adding bits and pieces to the master class files.
1999-03-19, Friday
I brought the spare CD drive along and plugged it in, using the same connector as the CD-R used. I leave the CD-R in but disconnected and I don’t connect the sound cable because I intend to keep the CD-R in Yaku. I’ll take out the spare CD after everything is installed.
Installation proceeds normal but afterwards, when actually starting the Startup Splash screen hangs. Indefinitely. The mouse and keyboard don’t react so I cannot do a ctrl+alt+del, I must reach for the reset switch. Booting in save mode gives a messages that you should start normal for the first time. Nice advice, normal doesn’t start. After trudling along for more than 10 minutes the desktop appears. In control panel there are no conflicts detected, everything looks and acts normal. There is however no CD and only 16 colours. Shutdown and restart gets, after about 5 minutes, back to the locked-up Windows98 logo. This is bullshit.
1999-03-20, Saturday
This weekend was reserved for resting.
Although ...
I went shoping and picked up a new mainboard, a soyo super-7 with 1MB of cach memory. I brought along an AMD K6-2 300MHz and a 64MB DIMM. I also braught home a cheep and cheerful Chianti, ideal for ad hoc drinking.
I did some cleaning up in the garden. It is not a big garden but it does need a lot of cleaning.
||
1999-03-21, Saturday
I made soup (leeks), took a bath and rested.
The computer was off the whole weekend!
||
Svenson © 1999
A day you don't learn something new is a wasted day.