Dear all,

this problem has been mentioned before by McDowell@vims.ed, but there was no solution so far posted. 
I'm using the current cloudman on Amazon AWS. When I download a fastq.gz file via an EBI-SRA ftp link using the "Upload File" function, although the file itself is loading to the /mnt/galaxy drive, at the same time /dev/xvda1 is filling up and at some point the download stops and says "[Errno 28] No space left on device". 
I've checked in the universe_wsgi.ini and the ftp upload path is set correctly to /mnt/galaxy/tmp/ftp.

Here's the df output:
Filesystem     1K-blocks      Used Available Use% Mounted on
/dev/xvda1      10321208   9797012         0 100% /
udev             8736124         8   8736116   1% /dev
tmpfs            3497960       232   3497728   1% /run
none                5120         0      5120   0% /run/lock
none             8744896         0   8744896   0% /run/shm
cgroup           8744896         0   8744896   0% /sys/fs/cgroup
/dev/xvdb      423135208    207440 401433728   1% /mnt
/dev/xvdf      104847360  22055572  82791788  22% /mnt/galaxy
/dev/xvdg      733900800 685188288  48712512  94% /mnt/galaxyIndices

I'm also having another issue: when I delete files in the history in galaxy, they don't acually get deleted in the files/000/ folder. And galaxy is still showing them under "Using ..." in the total space used. So if I download a 3GB fastq file, then delete everything in the history, Using is still showing 3GB and the file is still there in the folder. 

Could the two issues be connected? Are there maybe some rights missing for the galaxy instance to delete files?

Thanks in advance for any help
Best
Stanislas