Hi Everyone,
HPUX 11i running Universe 9.6 - does anyone know if there is a 4GB file
limit?
We have 64 bit files on a 64 bit file system, but one of these files grew
to over 4gb and blew up with an Internal Data Error on a BASIC write. I
thought that 64 bit files were supposed to be vv large...
Unix limitation? Perhaps something needs tweaking there.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
[EMAIL PROTECTED]
Sent: 01 September 2004 10:18
To: [EMAIL PROTECTED]
Subject: [U2] Memo: Universe file size limit
Hi Everyone,
HPUX 11i running
Surely if the standard Universe 32bit max file size is 2gb (2147483648 Bytes)
Then the 64bit would be 4gb (2147483648*2 Bytes)
Bob
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Adrian Matthews
Sent: 01 September 2004 11:46
To: [EMAIL PROTECTED]
Subject:
I forgot to mention - the FILEVAR will be considered to be ASSIGNED=TRUE
even if you've simply done a FILEVAR= somewhere in the program. The
key to this is never use the file handle variable name except for the
ASSIGNED or the various file-handling verbs.
Rick Ramsey
Healthpac Cumputer Systems
Hi Bob,
A 64bit file has a very large capacity. The 64 bit number is binary so
every bit = 1 greater power of 2. Whereas 2,147,483,648 = 2^31 then
9,223,372,036,854,775,808 = 2^63. A very big number indeed. The reason
it is 31 and 63 is that the leftmost bit generally signifies the sign of
the
Surely if the standard Universe 32bit max file size is 2gb
(2147483648 Bytes)
Then the 64bit would be 4gb (2147483648*2 Bytes)
I guess more in the neighbourhood of 2^63 = 9223372036854775808 bytes
but nevermind, the limit for a 32 bit system on 2gb is stupid as you could
have used the
I would check your ulimit settings.
On both HP-UX 10.20 and 11 we have a setting of 4194303.
Only problem with that is that IIRC this figure is in blocks - which would
mean 2GB.
Robert Paterson
Technology Support Manager
www.epicor.com
Tel.: +44 (1344) 468231
Cell: +44 (7799) 348513
Fax: +44
IF ASSIGNED(file.var) then
* Check for locks...
END
Works. You can also do NOT(ASSIGNED(file.var)) as well. This works on
both regular and file variables.
Hth
Colin Alfke
Calgary, Alberta
-Original Message-
From: Allen E. Elwood (CA) [mailto:[EMAIL PROTECTED]
Sent: Tuesday, August 31,
If I remember correctly, when I was implementing the 64-bit file system in
universe, the HPUX max file size was somewhere around 128 Terabytes, but I
can't locate my notes from that far back to verify that number. It
certainly sounds like something in the OS is stopping you, such as ulimit
Hi,
Thanks for all the responses.
Ulimit, is set to unlimited so that shouldn't be the problem. Our unix
support people cant find any other HPUX kernel parameter that would affect
the maximum file size. Glen's recollection is correct - HPUX support file
sizes up to 128 terabytes.
It is looking
The @select item should be a list of FIELD (i.e. dictionary entry) names not record
ID's
-original message-
[snip]
Below is the setup:
CT DICT BLAH @SELECT
@SELECT
0001 PH
0002 INTO ALTER - the records I want
Jeff Schasny | Denver, Colorado, USA | [EMAIL PROTECTED]
---
Hi Glen,
This is the output when we first ran across the problem:
0008 Program FIS00200: Line 551, Internal data error.
0009 Program FIS00200: Line 555, Internal data error.
0010 Program FIS00200: Line 555, WRITE failure.
Line 551 is a MATREADU from the file, and line 555 is a MATWRITE to the
Yes, this is the way it works, unfortunately. What we ended up doing
was to change att 8 in WO.COMP.QTY to
F;0(TWIP;X;;8);(G0 1);0(TWIP;X;;8);(G1 1);0(TWIP;X;;8);(G2
1);.;+;+;+;+ etc
-Dianne
Rod Hills wrote:
We are trying to do a very basic PICK technique of summing a multi-value
Well, I do know that the Internal data error message (at 9.6) is generated
within the file handling subsystem when blink values are mucked up;
interesting that no information pertaining to this type of error is
displayed so one might postulate this isn't the issue.
This message can also be
The translate returns MV data with spaces.
Dick Kryka
Director of Applications
CCCS of Greater Denver, Inc.
Paragon Financial Services
303-632-2226
[EMAIL PROTECTED]
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Rod Hills
Sent: Wednesday, September
Hi Rod,
I assist my customers who run both UniVerse and UniData in the Pick
flavor and created the following dict item to SUM the values from another
record:
0001: I (SQL)}}NUMERIC}Y}}}
0002: SUM(TRANS('WO-BOM',F0,3,'X'))
0003: MD4
0004: Test SUM
0005: 8R
0006: S
0007:
WO List
Hello!
We have had problems with big UV 64-bit files on HPUX11i PA-RISC.
As far as I remember, it is UV version 9.6.2.1 or 9.6.1.smth, where a
static hashed file
bigger than 4GB could be killed by modifying a record (beyond the 4GB
border?).
That is, if one is just adding records to a file it is
No. It has Itypes too, just that some of us love to code in a manner that
pre-dates Itypes. And for the challenge too! ;-) Oh. Or we have existing
stuff already coded to support.
At 12:40 PM 9/1/2004, you wrote:
Wow, in Unidata this would be an I descriptor that would just be
Dear All!
Please, shed some light on the strange behaviour of UV BASIC FMTS function.
Environment: UV 9.6.2.6 and 9.6.2.1 on HPUX 11i on PA-RISC 8xxx.
Problem: FMTS is supposed to FMT each field of a given dynamic array.
It appears to be wrong if source data in a field of the array contains
William,
Unless I missed it, there's no locational information in the post, which
makes
Relocation will not be considered - we are seeking a local candidate.
difficult.
--
- Charles Barouch
(718) 762-3884 x 1 - Key Ally Voice Mail
[EMAIL PROTECTED] - Consulting services
[EMAIL
Allen,
Universe does have I-descriptors, but the programmer was trying to build
an F-correlative (from old PICK) to do the task.
-- Rod Hills
-Original Message-
From: Allen E. Elwood (CA) [mailto:[EMAIL PROTECTED]
Sent: Wednesday, September 01, 2004 9:40 AM
To: [EMAIL PROTECTED]
Denver, Colorado isnt specific enough?
Jeff Schasny | Denver, Colorado, USA | [EMAIL PROTECTED]
---
u2-users mailing list
[EMAIL PROTECTED]
To unsubscribe please visit http://listserver.u2ug.org/
Oh gosh, read the subject? Denver CO sounds like a location to me :-)
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Results
Sent: Wednesday, September 01, 2004 10:59
To: [EMAIL PROTECTED]
Subject: Re: [U2] Senior Pick Programmer position Denver, Colorado
Hi Rod,
I too have had the 'pleasure' of F and A correlatives.
Once upon a time, a nice man named Dick Pick, who created Pick basic, issued
a challenge. He said that there was no way in native Pick to access the
last value of a multivalued field and challenged anyone to find a way.
Enter Me.
My with and without scores were 78 seconds and 80 seconds respectively.
Yes, without took longer.
It was 1000 updates, the same record set, and a simple RETURN after the
parameters. The trigger was AFTER INSERT UPDATE DELETE.
Maybe the INSERTS take longer or system load may have been radically
Problem: FMTS is supposed to FMT each field of a given dynamic array.
It appears to be wrong if source data in a field of the array contains
CHAR(247).
If you look at the DOCs for REMOVE (for example), it indicates that the
lowest system delimiter is CHAR(248), and that is true with REMOVE.
I think codingitis is the cause. Time to lay off the keyboard for a
while, man. You know, semi-colons start to disappear after 6 hours of
continuous coding. Any higher amount of time can cause strange things to
happen. I remember starting during mid-day and missing night-time once,
while
Since I haven't done this in some time. I cant remember how to do this. I
added a printer that I now wish to delete. I am using a dumb terminal to do
this via the SYSADM.MENU menu. There appears to be menu options at the top
File, Action, Help. I guess I need the magic keystroke for Action.
But Debra, all you need is an A below the B, on your (ahem, ahem),
and then you'll have a TBA tattoo (To Be Announced). You can have a
lucrative career doing commercials for TV Guide.
It was like experiencing a nuclear explosion in a very small place. -
Loni Anderson, describing having sex with
I've heard once that UniVerse goes down to char(246) as system delimiters.
Don't know where it came from. (Probably this list)
I would love to have the ability to access those values with the normal
tools of Basic or UniObjects, but I guess this is one case where there show
through.
Martin Scholl
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Tony Gravagno
I've tried the !NEWPCODE and !NEWVERSION ...
Sorry, I missed that detail in my previous response. The problem I
usually have w/ NEWPCODE is the syntax. You might try using FIBCDFN
in the AE editor and see if that
I'm no Wintegrate person, so I can't say what's happening there, but I
have seen terminal types send an ASCII 127 (DEL) for a BS. To the
terminal it looks perfectly normal but to the command interpreter it's
totally whacked. What does PTERM -DISPLAY show?
-Original Message-
From:
32 matches
Mail list logo