Fasterq-dump and fastq-dump failing on usegalaxy.eu with connection validation error (sra-download-internal)

Hello,

I am experiencing repeated failures when trying to download SRA accessions on usegalaxy.eu using both fastq-dump and fasterq-dump.

Example accession:

SRR2033948

The job repeatedly fails with the following error in stderr:

fasterq-dump.3.1.1 err: insp_get_file_size(
'https://sra-download-internal.ncbi.nlm.nih.gov/...SRR2033948.sralite.1',
remotely = YES ) -> RC(rcNS,rcNoTarg,rcValidating,rcConnection,rcNotFound)

connection not found while validating within network system module
- error with https open 'https://sra-download-internal.ncbi.nlm.nih.gov/...'

the input data is missing the QUALITY-column
fasterq-dump quit with error code 3

The job retries 3 times and fails consistently. No .fastq files are produced.

This happens with multiple SRR accessions, not only SRR2033948.

It appears that the failure occurs during HTTPS validation/open of the sra-download-internal.ncbi.nlm.nih.gov endpoint, suggesting a network or DNS connectivity issue between Galaxy worker nodes and NCBI.

Could you please confirm:

  1. Whether there is a known issue with SRA downloads on usegalaxy.eu?

  2. Whether outbound connectivity to NCBI SRA endpoints is currently restricted or unstable?

Thank you very much for your help.

Best regards

Welcome @OzGuo

These are strange results!

Sometimes the SRA can get busy at times and reject connections with odd reasons (everyone is impacted, not just Galaxy users).

To try to reproduce this, I’ve started up two tests in this history to retrieve these GEO fastq reads for → https://www.ncbi.nlm.nih.gov/sra/?term=SRR2033948

I’m 1) getting the reads directly with the Faster Download and Extract Reads in FASTQ format from NCBI SRA tool using the accession and also 2) pulled in the sra archive to get the reads that way instead (using the same tool to read the archive). Let’s see if I can reproduce the problem.

It seems that both of those methods worked now! Would you like to try again as well? I’ll leave this history shared in case you would like to examine the methods closer.

Please let us know if you are able to get a successful run or not! :slight_smile:

It works now for me as well. It looks like there was some problem at the NCBI’s end during the last 24 hours. Thank you @jennaj .

2 Likes

Thanks for letting us know! :rocket: