I'm reposting over here from the user side, since this is a local instance, and it was recommended. Hi all, We are running Galaxy on an Ubuntu 11.10 computer (5 TB, stripped, etc.). We are assembling a small genome (110 Gb). Our dataset isn't directly uploaded, but is accessed from a directory (if that matters). Everything went fine through the FASTQ Groomer, but when we ran Reverse-Compliment, we got the following error: fastx_reverse_complement: writing quality scores failed: File too large gzip: stdout: Broken pipe Any help that you might have would be greatly appreciated! Thanks! As a follow up, the file that we're trying to reverse compliment is ~26 Gb. The files seem to work fine until 2.1 Gb. There is plenty of memory (we have 3.4 Tb free on this system) and it doesn't seem to be an issue with the permissions or partitions. I also made sure that the Gzip and connectors to perl are up to date. And I have set everything in ulimit to be unlimited (so there are no issues for Ubuntu for the file size creation). This is a local instance, btw, though I'm sure that's obvious... We were able to groom the files to create the 26Gb file that we want to work with, so it seems like the computer should be able to do all of this... Any thoughts?