I am using script which reads some information from a text file and saves data into an embedded database, which I run from a linux shell. The text file contains 100 million entries so the script needs over 10 hours to complete.
The problem I have is that after running for 2-3 hours, the script almost stops: the process uses only 1% of the CPU and its progress virtually stops. I tried two different databases: Berkely DB and TokyoCabinet and the problem aoccurs with both. I do not think this is a buffering problem since I made sure the std output is unbuffered. What am I missing?
Thank you,
Radu