Hi y'all, I've been running into problems with Aspera seeming to have an incredibly high failure rate, leading to a bunch of a problems downloading a large study. Is this something that sounds familiar to anyone? (Running on a pool of Azure VMs) The symptom is a huge amount of retries. Many jobs are completing, but they seem to be biased towards smaller file sizes. When I investigate files on the nodes themselves, it seems as if at least some proportion of the failures are coming from MD5 mismatches in the downloaded files -- in the example below, I downloaded the md5sum and fastq.gz of that failed job from the VM filesystem directly, and indeed there's an MD5 mismatch.