Hi all,
I'm a little confused... does the following PDF not answer the UO.NET questions
posed so far?
http://publibfi.boulder.ibm.com/epubs/pdf/25124070.pdf
You can get to this link by going to the U2 Library:
http://www-306.ibm.com/software/data/u2/pubs/
selecting "Manual Online" which takes
>I hope somebody from IBM-U2 will post some release dates for UO.NET here.
>Otherwise wait until September :-)
The date is in the link that is part of the message you replied to.
Regards,
LeRoy F. Dreyfuss
Advanced Technical Services - U2 Technology Analyst
IBM U2 Data Management Solutions
Tel
UO.net will be included in the release of UD 6.1, which is due Aug. 10.
- Original Message -
From: "Rajesh Menon" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Tuesday, August 03, 2004 7:22 PM
Subject: RE: [U2] UniObjects for .NET (UO.NET)
> I hope somebody from IBM-U2 will post so
I hope somebody from IBM-U2 will post some release dates for UO.NET here.
Otherwise wait until September :-)
-Rajesh
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Tony Gravagno
Sent: Tuesday, August 03, 2004 3:03 PM
To: [EMAIL PROTECTED]
Subject: RE: [U
I have used nice for Unidata on HPUX many times and never had any problem.
Renice is the command to change nice level. Use "man renice" to get help on
the command. While using renice command just make sure to use correct pid
for the process.
Another command you may want to have a look is "kill"
>From the text on the website and the conference info below I don't see how
UO.NET is similar or different compared to UO except that it's re-written to
be managed code. Does it support the full UO API? Does it support ADO.NET?
Connection pooling is interesting. XML/DataSet support? I hope that
Thanks for the tips, Colin. I've added them to my collection of goodies.
Great point, too, about making sure nobody has the file open before deleting its
index. Not sure if the docs have been updated, but they didn't used to warn you about
this. So thinking it was safe, somebody here ran DELE
Ang Suan
Please note: in order to repair a 64 bit file you *MUST* use FIXTOOL -
uvfixfile cannot be used to repair a 64 bit file. You can use UVFIXFILE to
test of course - though there are better ways of doing this without (by
default) actually running UVFIXFILE over a complete file if it happens
> David T. Meeks
> The only real point of the UniJava class was to provide a centralized
> marshalling and management class for the API. It provided a mechanism
> to manage multiple sessions, etc...
I defer to the author. ;
So... Is the 81009 coming from trying to reconnect a UniSession that's
b
Dave,
I too will give this a shot.
Try a LIST.READU at the TCL level to see if your program is locked. If so, then
(again at the TCL level) try,
> RELEASE IBP AUTO.CAT.REQ
to see if the lock is released.
HTH!
Grant
-Original Message-
From: Thomas Derwin [mailto:[EMAIL PROTECTED]
Se
Dave,
I too will give this a "stab".
Try the following at TCL,
> RELEASE IBP
-Original Message-
From: Thomas Derwin [mailto:[EMAIL PROTECTED]
Sent: Tuesday, August 03, 2004 2:05 PM
To: [EMAIL PROTECTED]
Subject: Re: [U2] edit FI failed
Hi Dave,
Nobody seems to have responded, so I'
Hi Dave,
Nobody seems to have responded, so I'll give it a go.
My best guess is there's a file permissions issue. Are you sure you, or the user
you're logged on as, has write permissions on IBP?
I haven't seen an (E) option abort cause this problem on UD. Are you running UV?
Hope this helps,
SHM_FIL_CNT is the likely culprit.
You can monitor the use of this shared memory table that maintains the
current modulo of dynamic files with '$UDTBIN/sms -d | pg'
for example:
$ $UDTBIN/sms -d | pg
Device No. I-node No. FlagModulo Counter
11769475 377106
Actually, it wasn't for licensing. Licensing is still accounted for, but it's
dealt with on the server side.
The only real point of the UniJava class was to provide a centralized
marshalling and management class for the API. It provided a mechanism
to manage multiple sessions, etc...
I original
I think your assumption may be incorrect. I recently restored some indexes
into an account where the files weren't "defined" as having an index and it
caused some problems.
However, it is easy to trick UD with a bogus index. I just copy another X_
file and rename it. If one doesn't exist I simply
[EMAIL PROTECTED] wrote:
> I am writing some Uniobjects for Java code, and am having
> problems with the connect method.
> The first connect to the Universe system is
> successful, but if I then disconnect and then try to connect again, I
get the
> error[IBM U2][UniObjects for Java][UniSession E
FWIW, I learned the hard way that UD keeps a flag in the file header that knows
there's an index on the file, then looks for the index in the same directory.
Restored a file from a backup tape to another directory, but UD refused to let me even
read the data until I went back and restored the in
> FWIW, our newer disk cabinet with faster SCSI controllers and higher disk
RPM
> (but fewer spindles) blows away the older one that has 2-3 times the
spindles. :-)
> but 36 GB drives that cost a lot less, and that are a lot faster, and last
a lot longer
It's actually hard to think that a shop w
What happens if you reinstantiate the session object with:
uSession = uj_sess.openSession();
instead of:
uSession.connect();
?
HTH,
David Beahm
[EMAIL PROTECTED] wrote:
Hi everyone,
I am writing some Uniobjects for Java code, and am having problems with th
e
connect method. The first connect to th
Hi everyone,
I am writing some Uniobjects for Java code, and am having problems with the
connect method. The first connect to the Universe system is successful, but
if I then disconnect and then try to connect again, I get the error[IBM
U2][UniObjects for Java][UniSession Exception][ErrorCode: 810
Gotta look at the disk specs. All things may not be so equal. Disks are getting
faster and smarter as they get larger.
Latest disk speeds are 15,000 RPM vs. 8k or 10k. Faster SCSI buses, bigger caches and
smarter access algorithms on both the drives and the controller cards boost I/O
through
>Subject: RE: [U2] many spindles - was Copying data between two UV servers
> It's true that it's getting tougher to justify 4 spindles that happen to
be 36gig each for a database that is only 15gig in total size. But the
speed difference that is achieved by splitting the data files between 2 set
I don't know if anyone replied yet, but I will try and answer this.
Depending on your unix flavor, the nice values will be different. Linux
nice values are between -20 (highest priority) and 19 (lowest priority),
with a default of 0. For hpux, the values are from 0 to 39, with a default
of 20. I
It's true that it's getting tougher to justify 4 spindles that happen to be 36gig each
for a database that is only 15gig in total size. But the speed difference that is
achieved by splitting the data files between 2 sets of mirrored disks is something
that can be easily experienced by simply ru
I will be out of the office starting 07/31/2004 and will not return until
08/16/2004.
I will respond to your message as soon as possible.
---
u2-users mailing list
[EMAIL PROTECTED]
To unsubscribe please visit http://listserver.u2ug.org/
> It's not just U2 that likes to have multiple spindles.
> Any disk intensive processing activity will benefit by having the load
split accross multiple disk subsystems.
Agreed .. So why when we bought our new RS6000 about 1 year ago did the IBM
guys try to sell us on a SMALLER number of LARGER di
All,
Please try to trim unneeded quotes from the bottom of your emails,
whatever you requote is requoted in the digest, making the entries too
large for some users.
- Charles Barouch, Moderator
---
u2-users mailing list
[EMAIL PROTECTED]
To unsubscribe please visit http://listserver.u2
At 07:55 PM 8/2/2004, you wrote:
No, UniData does not have this problem because UniData does not store a path
to the index file in the header of the data portion. On OPEN, UniData looks
for correctly named index files (X_file for static files and idx00n for
dynamic files) in the SAME directory as
Thanks Jonathan,
I actually ordered an upgrade about 2 months ago. Our VAR rarely responds with a first
request... Having worked for a couple different VARs during the past decade and a
half, I'm hyper-critical of unresponsiveness.
Karl
On Tue, 3 Aug 2004 09:51:50 +0100
Jonathan D Smith <[EMAI
It's not just U2 that likes to have multiple spindles. Any disk intensive processing
activity will benefit by having the load split accross multiple disk subsystems. This
goes for Oracle, DB2, Microsoft SQL Server, UniVerse, email systems, etc, etc.
Think about it, if your writing an index on
> In unix, use "soft links". See man pages on "ln -s".
> You can do that today to spread out INDEX.000, INDEX.001 INDEX.002, etc.
> across many spindles.
A few comments:
First, using a symbolic link (ln -s) will cause every OPEN to have to read
twice .. Once to read the link, and once to read th
Dear David,
Yes, i got it. Thanks for your information
Thanks and Regards
> -Original Message-
> From: Logan, David (SST - Adelaide) [SMTP:[EMAIL PROTECTED]
> Sent: Tuesday, 03 August, 2004 4:06 PM
> To: [EMAIL PROTECTED]
> Subject: RE: [U2] FW: How do I determine whether
> > The way I really hoped would be implemented was to have the index
> > files exist within the file subdirectory, i.e. all files with
indexes
> > would be similar in structure to Type30, but have a .index
> > subdirectory within them. So, from then on, the indexes would be
> > "tied" to the m
Does anyone have any experience of what the "nice" and "priority" values
should be and how they can be changed?
--
* This message has been scanned for viruses and dangerous content by
* Blairs of Scotland MailScanne
Try it like this:
0001: RTVEC = ""
0002: VNM = TRIM(FIELD(RTVEC<8>,":",2))
0003: END
Bottom at line 3.
: .X BAS
"TEST" filed in file "BP".
Executing the command "BASIC BP TEST".
Compiling: Source = 'BP/TEST', Object = 'BP.O/TEST'
Compilation Complete.
Bob
-Original Message-
Should this not read:
VNM = TRIM(FIELD(RTVEC<8>,":",2))
NOT
VNM = TRIM(FIELD(RTVEC<8>),":",2)
or am I missing something?
Regards
Bjvrn Behr
Programmer
HYFLO Southern Africa (Pty) Ltd
Tel : +27 11 386 5800
Fax : +27 11 44
Hi Karl,
This problem does not happen on UniVerse 10.1.4 on Aix See Screenshot
below :-
001: RTVEC = ""
002: VNM = TRIM(FIELD(RTVEC<8>),":",2)
003: END
Bottom.
*--: FIB
Filed "TESTER" in file "BP" unchanged.
Compiling: Source = 'BP/TESTER', Object = 'BP.O/TESTER'
02VNM
Hi Ang,
To change to a 64bit file just put 64BIT on the end of the resize
command. The documentation is in the USER REFERENCE book. The manuals
are available at http://www-306.ibm.com/software/data/u2/pubs/library/
In order to tell a 32bit file from a 64bit file, you can use filepeek
(this is in
Dear Steve Ferries,
Hi . Just wondering if you resize the file from 32 to 64 bits using
RESIZE command ? Coz i having a problem when using RESIZE at Universe
version 10. Do u have any idea to change the file from 32 bit to 64bits ?
error messages is below during resize
mkdbfile: unable to
39 matches
Mail list logo
