Failed to download pb-assembly in conda

Question

[Maliha]

I am trying to download pb-assembly in conda. In their installation guide in GitHub it is said to follow this path:

conda create -n denovo_asm
source activate denovo_asm
conda install pb-assembly

However after running the "conda install pb-assembly" I am getting this following error everytime

Collecting package metadata (current_repodata.json): done Solving environment: failed with initial frozen solve. Retrying with flexible solve. Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source. Collecting package metadata (repodata.json): done Solving environment: \ Killed

I also tried to download this tool using this command line:

conda install -c bioconda pb-assembly
conda install -c "bioconda/label/cf201901" pb-assembly

But still I am getting the same error. I'd be grateful if anyone can help me to solve this.

[Jesse]

Wait, Killed? What do you see if you run echo $? immediately after the conda install fails? I'm wondering if it's running out of memory while trying to do the package dependency resolution. If this is on Linux you could probably see a block of text about this in sudo dmesg output, with lines labeled "OOM" (out of memory) if my random guess is correct.

[Maliha]

echo $? gave me 137 after running it

[Billy Rowell]

Good call @jesse. Looks like conda is running out of memory. What about trying mamba instead of conda.

conda install mamba
...
mamba install pb-assembly

[Jesse]

Yep! (FYI Maliha, the 137 from $? is the exit code of the conda process, where 137 = 128 (exit code is due to signal received by the program) + 9 (the signal was SIGKILL), which almost certainly implies the OS sent a kill signal due to memory usage.)

I'd strongly second the mamba suggestion. I've had it help in exactly this situation (trying to use conda on a VM with "only" 1 GB RAM and running out) and on top of that it tends to be much faster than conda. It should do the same exact thing from your perspective, but internally it figures out package dependencies with its own method, different from conda's.

[Maliha]

Thank you for your replies. I have tried to install pb-assembly with mamba install pb-assembly but I am getting this error

Encountered problems while solving: -package pb-assembly-0.0.7-0 requires pb-falcon >=0.3.0, but none of the providers can be installed

But In GitHub, they wrote that with conda install pb-assembly all the dependencies would be automatically installed. So, Idk what should I do next.

[Jesse]

Hm, works for me-- it finds pb-assembly 0.0.8 which pulls in pb-falcon 2.2.4. What channels are you using?

Maybe one thing to note: the mamba docs say that mamba needs to be in the base environment ("installing mamba into any other environment than base can cause unexpected behavior"), whereas you'll want pb-assembly in its own environment like you were trying. So if you were starting from base, it might be something like:

$ conda install mamba
$ conda create -n denovo_asm
$ conda activate denovo_asm
$ mamba install -c bioconda pb-assembly

...making sure you see (base) before the shell prompt for the first command. Otherwise you might want to remove any per-env mamba and make sure it's in the base environment instead.

[Maliha]

Thank you so much Jesse and William for helping me out. I installed pb-assembly now.

Answer 1

[colindaven]

Try removing this env, then creating a completely new env again. Also, it may take some (or a long) time.

The tool mamba will make your installations faster as well.

[Maliha]

Do you mean I should remove the denovo_asm environment? I have already tried that. But I am getting the same error.

Answer 2

[Billy Rowell]

I wonder if this is due to the change to strict channel priority in conda.

Could you try:

conda config --set channel_priority flexible
conda install pb-assembly

If this works, please create a GitHub Issue.