hopefully a not stupid question.. was working on a lsf cluster i have access to, and wanted to use fetchngs. it produced errors that i thought suggest network troubles, namely failing to resolve a reference to an ftp site, and wget failing w an exit code of 4. but if i try any of the individual wget commands from the logs outside the pipeline, it works. i was hoping someone would have an idea why that would be..

I think I’ve seen the same problem with v1.12+ (LSF + Singularity). I never took time to confirm and investigate / raise a bug. Still using v1.11 at the moment

well good to know 1.11 works for you under similar circumstances. can confirm ive been trying 1.12

@Mahesh Binzer-Panchal. you just reported a fetchngs error in https://nfcore.slack.com/archives/C07ACNLELR2/p1720710487780759?thread_ts=1720701418.546999&cid=C07ACNLELR2 . I wonder if it’s the same error we’ve faced here

Maybe. I can pull from outside it seems. I'll investigate tomorrow though.

download method aspera also works fine.. and 1.11 was using curl rather than wget it seems. bad wget container?

Maybe outdated certificates. Thanks for looking into this and figuring out it works on dev.

A long time ago, there used to be busybox images used as base for some tools that broke name resolution with singularity (not certain I remember the details but think they had

/etc/resolv.conf

as a link to

/var/tmp/something

which was in an area singularity mounted and it then didn't mount the host

resolv.conf

over the one in the image). But I think that was fixed ages ago.

I’m having the same issue, but I think our compute nodes don’t have “internet” access, just internal network access.