I wanted to download the accepted junction .bam file from tophat output of my local instance and I get an "out of memory" error. When I examine the server via command line, I found that a python process used by galaxy occupied more than 80% of total memory (on the virtual machine with 10G of RAM).. I tried curl command to retrieve the datafile after rebooted the virtual machine, and python is activated again and used up all the memory.

The bam is around 20G of size, but I never had this kind of problem with other tophat analyses before on my local instance although they are of the same size. The discription on the web mentioned some .dat files that I manage to find on the disk, but not the bam.

Can anyone explain what python is doing and how can I solve this please?

Thanks,

Delong