the record is too big to fit into block

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

thumati.praveen
Participant
Posts: 106
Joined: Wed Oct 04, 2006 5:21 am

the record is too big to fit into block

Post by thumati.praveen »

Hi,

Can you aware of below problem earlier ?


APT_CombinedOperatorController(5),0: Fatal Error: File data set, file "/etl/IS/Datasets/ds_isap_irde020_am_gm_gbo_final.ds".; output of "cpyROw": the record is too big to fit in a block; the length requested is: 145247, the max block length is: 131072.


Parameters for the job is:

$APT_CONFIG_FILE = /opt/IBM/InformationServer/Server/Configurations/default.apt
$APT_RECORD_COUNTS = False
$APT_DUMP_SCORE = False
$APT_MONITOR_SIZE = 500000
$APT_MONITOR_TIME = 20
$APT_DEFAULT_TRANSPORT_BLOCK_SIZE = 78904532


My input record length is 4000033960.


Thanks,
Praveen.
nagarjuna
Premium Member
Premium Member
Posts: 533
Joined: Fri Jun 27, 2008 9:11 pm
Location: Chicago

Post by nagarjuna »

Try to increase the value of the variable APT_DEFAULT_TRANSPORT_BLOCK_SIZE
Nag
thumati.praveen
Participant
Posts: 106
Joined: Wed Oct 04, 2006 5:21 am

Post by thumati.praveen »

I was already specified 8 digit number still job looking 131072 this.
ArndW
Participant
Posts: 16318
Joined: Tue Nov 16, 2004 9:08 am
Location: Germany
Contact:

Post by ArndW »

I seem to recall that the block size is allocated as a power of 2 and rounded from the number specified in the block size. Try specifying 1048576
thumati.praveen
Participant
Posts: 106
Joined: Wed Oct 04, 2006 5:21 am

Post by thumati.praveen »

ArndW wrote:I seem to recall that the block size is allocated as a power of 2 and rounded from the number specified in the block size. Try specifying 1048576 ...
I amgetting same problem.

Even I given more then that what you have specified.

Thanks,
Praveen.
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

Your input record length is over 4 billion bytes? :?
-craig

"You can never have too many knives" -- Logan Nine Fingers
thumati.praveen
Participant
Posts: 106
Joined: Wed Oct 04, 2006 5:21 am

Post by thumati.praveen »

chulett wrote:Your input record length is over 4 billion bytes? :?
Yes boss,You identified my actual problem.

Could you please advise how to resolve the problem?

when ever small amount of records comming means less than 50000 records it is working if the records are comming more than 100000 I am facing this probel.

Thanks,
Praveen.
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

I have no idea how to handle something that big, sorry. Hopefully someone else does.
-craig

"You can never have too many knives" -- Logan Nine Fingers
thumati.praveen
Participant
Posts: 106
Joined: Wed Oct 04, 2006 5:21 am

Post by thumati.praveen »

My input reocord is combination of xml chunks.

So many xml chunks I am taking as a single column in my input row.

Thanks,
Praveen.
ArndW
Participant
Posts: 16318
Joined: Tue Nov 16, 2004 9:08 am
Location: Germany
Contact:

Post by ArndW »

If you are getting the same error message, then obviously the change isn't being accepted. Can you check the value in the director log for your run?
thumati.praveen
Participant
Posts: 106
Joined: Wed Oct 04, 2006 5:21 am

Post by thumati.praveen »

ArndW wrote:If you are getting the same error message, then obviously the change isn't being accepted. Can you check the value in the director log for your run? ...
I checked the logs in the datastage.It is taking what ever values we are passing.But when it reaches the 22389 record count it is failing the job with above reason.

I am still afrid why it is not taking values from APT_DEFAULT_TRANSPORT_BLOCK _SIZE.

could you let me know is their any other settings I need to specify in the job?

I am assuming mainly the problem occures while writing the output dataset on the disk.

Thanks,
Praveen.
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

Just how same is the "same error" you get after changing the APT variable? Exactly the same as in it still says "the max block length is: 131072" or just that it is still "too big" but you see the "max" changing?
-craig

"You can never have too many knives" -- Logan Nine Fingers
thumati.praveen
Participant
Posts: 106
Joined: Wed Oct 04, 2006 5:21 am

Post by thumati.praveen »

chulett wrote:Just how same is the "same error" you get after changing the APT variable? Exactly the same as in it still says "the max block length is: 131072" or just that it is still "too big" but you see ...
It always gives below error:

APT_CombinedOperatorController(4),1: Fatal Error: File data set, file "/etl/IS/Datasets/ds_isap_irde020_am_gm_gbo_final.ds".; output of "cpyROw": the record is too big to fit in a block; the length requested is: 145247, the max block length is: 131072.

what ever value you set to Environment variable the error is same.

Thanks,
Praveen.
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

There are other APT variables that may help, check this post to see if anything there works for you.
-craig

"You can never have too many knives" -- Logan Nine Fingers
ArndW
Participant
Posts: 16318
Joined: Tue Nov 16, 2004 9:08 am
Location: Germany
Contact:

Post by ArndW »

Are you just setting this in the ADMIN or are you also passing it to the job as a parameter (which is what you should be doing).
Post Reply