Idiosyncratic Local Docker Setup Issue

Hello! I recently downloaded the NCBI PGAP software onto my SSH server and am having issues with Docker. The PGAP software is for genome annotation, and in the installation package, it comes with a test genome to see if the software downloaded correctly. I was successfully able to run the program with the test genome, but when uploading my own genomes to my server, I got an immediate failure. I have been talking with some support people for PGAP, and they concluded it’s likely due to an idiosyncratic local Docker setup issue, because I am able to ls the file, but not cat the file. Here’s our full conversation if you would like.

Can you share the ouput of the ls command test? Who is the owner? When you can run ls that just means you could list the files in the folder. So you had read permission to the folder, but if your user (UID 1000/GID 1000) can’t read the file, cat won’t work.

This is the output:

-rw-rw-r-- 1 gpipe gpipe 9 Feb 28 19:04 /pgap/user_input/GCA_019429385.1_ASM1942938v1_genomic.fna

I am the owner as well.

Edit: Just compared the file sizes of the genome already installed and the genome I uploaded, here is the genome already installed:

-rw-r--r-- 1 ubuntu ubuntu 588482 Oct  4 15:53 ASM2732v1.annotation.nucleotide.1.fasta

and then the one I uploaded:

-rw-rw-r-- 1 ubuntu ubuntu    9 Feb 28 19:04 GCA_019429385.1_ASM1942938v1_genomic.fna

The file sizes on Github are correct (30,110 for the GCA file).

I’m trying to follow the the topic, so you uploaded a file to your server which has an incorrect size on the server as well?

The ID is the important part, not the name, but since everyone should be able to read the file, it should work.

My question is now, how did you install Docker on what OS?

Yes, I can see the correct file size on github, but when I use curl to bring the file onto my server, the file size is not nearly as big as it should be.

I installed Docker through ubuntu on my server.

I’m not sure how the issue is related if the file on the host is not correct either. If the file is small, that is probably because there the url was wrong or there was a redirection and you downloaded an html. On the other hand, if that were the case, you would be able to read the content. Check it on the host. I usually use the “less” command that warns me if the content is binary and shows me the content if it is text.