SRAtoolkit: error while running fastq-dump

Question

[arnstrm]

Hi,

I am trying to download few SRA files from NCBI SRA database using SRAtoolkit. The command I am running is

fastq-dump --origfmt --gzip --split-files --accession SRR447946

But after a while of running (downloading), it gives an error:

=============================================================
An error occurred during processing.
A report was generated into the file '/home/arnstrm/ncbi_error_report.xml'.
If the problem persists, you may consider sending the file
to '[e-mail]' for assistance.
=============================================================

2014-10-20T16:04:57 fastq-dump.2.3.4 err: file descriptor invalid while constructing file within file system module - curl_easy_perform( FileRead, 1057095680.131072 ) failed with curl-error 'CURLE_COULDNT_CONNECT' (7)
2014-10-20T16:05:08 fastq-dump.2.3.4 err: file descriptor invalid while constructing file within file system module - curl_easy_perform( FileRead, 234749952.131072 ) failed with curl-error 'CURLE_OPERATION_TIMEDOUT' (28)
2014-10-20T16:05:08 fastq-dump.2.3.4 err: buffer insufficient while executing function within transform module - failed SRR447946

Does anyone know what is the problem here? Quick Google search didn't find any solutions.

Thanks for any help!

Answer 1

[Renesh]

You are getting error CURLE_COULDNT_CONNECT; it means your connection to NCBI server is failed. Check you connection and port. retry the download. When I tried on my system, it is working fine.

[Renesh]

If you still getting the error, download those manually by going to SRA site http://www.ncbi.nlm.nih.gov/sra/?term=SRR447946

[arnstrm]

It is on HPC. So you think the compute nodes might be having connection issues? I haven't had problems downloading the files before, but I will write to the sysadmin to see if they can solve this problem.

Thanks for the tip!

[Renesh]

On HPC, this should work;

wget ftp://ftp-trace.ncbi.nlm.nih.gov/sra/sra-instant/reads/ByRun/sra/SRR/SRR447/SRR447946/SRR447946.sra

[Devon Ryan]

It's quite possible that the worker nodes are walled off from connecting to the rest of the net. Use wget as Renesh said and then use fastq-dump and you should then get the files you want.

BTW, you can save time and just get this dataset from ENA. Then you don't have to deal with the annoyance (and time) of converting to fastq.

[arnstrm]

Great, thanks for EBI link. I will try that!