getrandom error with pod5 convert fast5

[cocvincent]

Hello,

I have a problem when trying to convert fast5 files to pod5 files with pod5 convert fast5 (pod5 v0.3.6).

In both cases, I get this error in the log file:

The job is not killed when the error occurs.
I'm using 100G of memory with 50 cpu on a computing cluster.

I hope you can help me, thanks.

[HalfPhoton]

hi @cocvincent,
Can you add this environment variable before your command to generate log files please?
Please also reduce the number of threads to 2 to generate fewer log files
POD5_DEBUG=1 pod5 convert fast5 -t 2 --output outDir/ ...

and then share the .log files generated here?

Kind regards,
Rich

[cocvincent]

Hello,
Thank you for your quick reply !

Here is the command:

And here 4 .log files generated:

2024-03-01--14-02-51-main-pod5.log

Sorry, the last file is too big for a simple screenshot.

[HalfPhoton]

Hi @cocvincent - can you try and run pod5 in a normal python virtual environment please? We've seen a few issues recently with users running in conda.

# Create a new virtual environment called "venv"
python3.12 -m venv venv --prompt=pod5
# Activate the venv
source venv/bin/activate
# update pip and install pod5
pip install -U pip pod5==0.3.6
# run your command
POD5_DEBUG=1 pod5 convert fast5 ...

Kind regards,
Rich

[HalfPhoton]

@cocvincent, do you have any updates on this issue?

[cocvincent]

Hello @HalfPhoton ,
I have forwarded the request for installation on the cluster.
I'm waiting for a reply and will keep you informed.

Thanks !

[cocvincent]

Hello,
The version of Python on our computing cluster is too old to install pod5 directly.
Is using a conda environment and installing pod5 via pip a viable option?

Thanks

[HalfPhoton]

Potentially. I suggest giving it a try.

Can your cluster administrators update python? If your cluster doesn't have support for python3.8+ then many things will slowly stop working as 3.7 had it's end of life last October.

Kind regards,
Rich

[cocvincent]

Hello,

The cluster administrator installed python 3.12 and then pod5 with PIP.
I tried:

and got 4 outputs:



2024-03-15--08-27-54-main-pod5.log

It's very similar to what I've had before.

Thank you for your time.

[HalfPhoton]

Thanks for the update - we'll take a look at this and get back to you.

[HalfPhoton]

@cocvincent - Can you share some information on your system OS etc?

[cocvincent]

Hello,

here's the information on OS:
cat /etc/os-release

NAME="SLES"
VERSION="12-SP1"
VERSION_ID="12.1"
PRETTY_NAME="SUSE Linux Enterprise Server 12 SP1"
ID="sles"
ANSI_COLOR="0;32"
CPE_NAME="cpe:/o:suse:sles:12:sp1"

Unfortunately, it's not possible to update the OS version.

Is there a docker image available?

Thanks.

[HalfPhoton]

Could you tell me your glibc version?

[cocvincent]

The glibc version is 2.19

[HalfPhoton]

Ok - this reminded me of this similar issue
#63

The glibc version you have is too old: #63 (comment)

Are you able to compile the project yourself? https://pod5-file-format.readthedocs.io/en/latest/docs/install.html#installation-from-source