memory error 1

[wernelr]

Hi,

Can anyone help me with this error: "not enough memory to execute map"?

Aside from closing all other applications running on Windows, are there parameters that can be set in the Mercator application environment setup?

Any help will be appreciated.

Thanks,

 

[janhes]

Have you set your workfiles to memory? If so change it to file. You could also try running with your input card in burst mode.

 

[wernelr]

Thanks for the quick reply. I tried running in burst mode but the results are the same.

I also tried having workfiles set to memory (they were initially set to file by default) but I still get the same message.

I am a bit confused since my input file is a measly 5MB of data.

Do you have other suggestions? It used to work perfectly okay.

BTW, I am working on an SAP BAPI.

Regards,

 

[janhes]

Can you give us some more info. What version of mercator, are you running under event server or command server, what platform, has there been any change to the data?
I have seen this problem where it tried to read all the data in as a single item - no terminators found.
Are you using a trace file and is there anything usefull in there?
Jan

 

[wernelr]

hi janhes,

Thanks again for the interest. We're using mercator version 6.5 and I'm running the map in question under a command server using windows 2000. No, there has not been any change to the data and a trace file was not generated even if specified in the settings.

I've created a new map to do the same thing, but instead of using the SAP R/3 BAPI adapter, I decided to read files directly from the database server (informix) using the database adapter.

It ran okay on a Sunday night when no one else is working but won't run on a Monday with other people logged in the network. This is weird because I don't think network traffic affected the map processing.

Have you or anybody else got a clue?

Regards,

 

[janhes]

I suspect the database is locking you out. What error dis you get on the Monday?
The fact that no trace file is produced even though it is set is strange. I have only seen that happen when it could not identify any objects within the input data and the size of the data caused the map to crash.
Did you get a trace file when you ran it on Sunday?
You could try posting your map, type tree, logs and any trace and sample data to

http://www.eaiusergroup.com

and I'll have a look at it.
Jan

 

[wernelr]

hi janhes,

I sent an attachment for you in the site you suggested inside the middleware -> mercator forum. Hope you find it in order.

regards,

wernelr

 

[janhes]

Hi wernelr,
I have downloaded your file and I have managed to run it withour any problems on a Win NT PC using 6.5.2 both in burst and integral mode. I also increased the size of the input file to 1mb and it was still ok.
However, there was one strange thing - every time I built the map and then ran it, it reported that the source file had changed since the last compilation. This is not correct behaviour.
I copied your map into a new one and it ran correctly. I'm wondering if there could be a bug at 6.5. I know I found several at that version and would suggest you upgrade to 6.5.2.
I will attach my source and complied map to eaiusergroup and you can give it a try.
Jan

 

[janhes]

I dont seem to be able to get the attachment to work. however, I would try creating a new map source to rectify the problem.
Jan

 

[janhes]

File attached now.
Jan