mobilink server crashes on work on a download_cursor of a big table

[micha_rath]

hi all,
here is the situation:
We using ASA 12.0.0.1483 with mobilink.
I have prepared and deployed my consolidated database and the created
model.
Set up a mobilink user and an agent, created some task to transfer the
remote database.
Agent is configured on the remote site and start to search for tasks.
mlfiletransfer transfers the deployed db files (.db .log) and starts
the databaseserver and the sync.
After sync is started and schema was successfully applied, mlserver
try to load the data of consolidated database.
If the database is small ( < 2 GB) transfer is done in a few minutes.
Now I try to set up mobilink with a database > 30 GB. Mobilink-Server
tries to cache all data. If cache is full it will swap into a file
(_T2) at temp directory. If file _T2 reaches 25GB (after 2.5h of work)
mlserver process crashing and writes some dump files. No errors
reported to mlserver logfile. On client site the transfer also closed
when mlserver goes down.

here the comandlines:
mlserver12.exe -c "dsn=mlzentrale;uid=***;pwd=***" -o "staffxp-ml.txt"
-x tcpip -dl -vhmnos -ftr "PathToTransferDir" -ftru
"PathToTransferDir"
dbmlsync.exe -c "UID=***;PWD=***;ENG=**" -u User1 -s PublicationName -
v -o "StaffXP-ML.txt" -e
"ctp=tcpip;adr='host=192.168.2.121;port=2439';sch=EVERY:
00:10;nss=on;bd=off"

Is it a configuration failure?
Why mlserver tries to load all data at first?
How to set up a transfer order? Is it possible to transfer the tables
one after another?