Partition of Scratch Dbset full (FATAL MESSAGE 1221) 2

[Jagon]

Hi,

When I perform an analyse, I get this message :

USER FATAL MESSAGE 1221 (GALLOC)
THE PARTITION OF THE SCRATCH DBSET USED FOR DMAP-SCRATCH DATABLOCKS IS FULL.

I really don't understand how to extend the size (I have more than 100 Go free on the partition but Femap doesn't seem to "see" them).

Has anyone ever experienced this problem ?

Julien

 

[ThomasH]

Hi

I don't thing the problem is in Femap. I think its the solver (Nastran). What "flavour" of Nastran do you use (MSC/NX/NEi)?

Regards

Thomas

 

[Jagon]

I'm using Nx Nastran.

Thomas, thanks again for helping me : I see your answering on all my threads !

Regards,

Julien

 

[ThomasH]

Hi Julien

I have never used NX but I previously used MSC.Nastran (today it's NEi).

But, NX and MSC are (or at least were) very similair.

There is a setting where you can allocate memory in MSC and it should be the same in NX.

I've found an old tread regarding this:

http://www.eng-tips.com/viewthread.cfm?qid=98770&page=12

I hope it works.

Thomas

 

[Jagon]

Thanks,

Just to be sure :

Where do I have to paste this ?
Do I paste even the $ ?
What do I have to change if my scratch file is in the default location ?

Regards,

Julien

 

[saberfem]

Jagon,
If you are still using all of the installation defaults, then try the following:
In the Analysis Set Manager,On the NASTRAN Executive and Solution Options form, in the box labeled "system cells" type in the following:
buffsize=65537

 

[ThomasH]

Jagon

Now I'm at the office and I have picked the following from an old MSC-file:

$INIT MASTER(S)
$NASTRAN SYSTEM (151)=1
INIT SCRATCH LOGICAL=(SCR(70GB)), SCR300=(SCR300(70GB))
INIT DBALL LOGICAL=(DBALL(70GB))
ASSIGN SCR = './tmp.scr',TEMP
ASSIGN SCR300 = './tmp.scr300',TEMP
ASSIGN DBALL = './tmp.DBALL',TEMP

$NASTRAN OP2NEW=0
$ID D,MSC
$SOL SEMODES
$TIME 10000
$CEND

The lines with $ is just to orientate you in the dat-file. These are the very first lines in the file.

Since you run NX it might not work, but it also might work.

I hope it does

Thomas

 

[Jagon]

Thank you Thomas.
It seems to work.
I will confirm when calculation is done.

Regards,

Julien

 

[Jagon]

I processed the analysis which failed because of "EXCESSIVE PIVOT RATIOS IN MATRIX KLL".
Used to that, I fixed it and processed the analysis again.

And then, in the analysis monitor, after 5 to 20 seconds I get the following :

(...)
14:31:19 Beginning analysis
14:31:19 NSINIT: *** NSINIT ERROR 70 *** UNABLE TO OPEN UNIT 8
14:31:19 STOP
14:31:19 Analysis complete 3
NSINIT/NSCLOSE
(...)

And I get this again and again, even with the previous model (the one with the excessive pivot ratio)

It seems that Femap is the first soft that has an April Fools' Day plugin.

What do you think about that ?

Julien

 

[Jagon]

And I also tried to change the buffsize, in vain.

 

[ThomasH]

Hi

I think that "OPEN UNIT 8" can be related to the disk. I admit that I don't know why there is a problem but related to your other post om issues with saving. Could it be so that your disk is not feeling well? Have you tried defragmenting it or running a diskcheck.

The problems might have that in common.

Also, since you mentioned the analysis monitor. Are you running Nastran from within Femap. If so, try running it outside Femap.

Regards

Thomas

 

[jimb57]

it's a bug in NX nastran, they reported it.

Good news is Nei have posted a month's lease on there website to use NEINastran for a month or whatever it takes to fix it. Could keep you going - neat marketing from someone at Noran to spot the problem and make the offer.

 

[Jagon]

This is true Jimb57,

watch this:

http://www.eng-tips.com/viewthread.cfm?qid=241596&page=1

April fools' day, I told you!

Regards

Julien