From 380adf083ad0ec8bcb2de0c585449fbe17edfef4 Mon Sep 17 00:00:00 2001 From: Barry digby Date: Wed, 28 Jul 2021 11:41:16 +0100 Subject: [PATCH] markdowwwwn --- docs/lugh.md | 36 +++++++++++++++++++++++++++--------- 1 file changed, 27 insertions(+), 9 deletions(-) diff --git a/docs/lugh.md b/docs/lugh.md index e89a80a..0a17272 100644 --- a/docs/lugh.md +++ b/docs/lugh.md @@ -1,19 +1,37 @@ # nf-core/configs: LUGH configuration -To use the lugh configuration profile with your pipeline, add `-profile lugh` to your `nextflow run` command. Please take care to use the `-bg` flag or run the job on a compute node. DO NOT RUN ANALYSES ON THE HEAD NODE. +Author: Barry Digby +Contact Info: barry.digby@nuigalway.ie +System Administrator: Chris Duke -The configuration file will load prerequisite modules for users, however it is entirely up to the user to have a functional version of nextflow installed in their path. +## Quick Start -Lugh has 3 queues: +To use the lugh configuration profile with your pipeline, add `-profile lugh` to your `nextflow run` command: - * `MSC`: compute0{1..3}, 16 cpus, 32GB memory (Please reserve for MSc students only) - * `normal`: compute{10..29}, 16 cpus, 64GB memory. - & `highmem`: compute0{4..9}, 32 cpus, 128GB memory. +```console +nextflow -bg run nf-core/rnaseq -profile test,nuig +``` -The configuration profile design is very simple. If your process exceeds 64GB memory or 16 cpus, it is sent to the `highmem` queue. Else, `normal`. +Please take care to use the `-bg` flag, or run the job on a compute node. -Whilst I (@BarryDigby) am happy to help people set up their nextflow pipelines, please note that I am not the system administrator. +:warning: DO NOT RUN ANALYSES ON THE HEAD NODE :warning: + +The configuration file will load prerequisite modules for users (`Java` & `Singularity`), however it is up to the user to have a functional version of `nextflow` installed in their path. Follow `nextflow` installation instructions at the following (link)[https://www.nextflow.io/docs/latest/getstarted.html#installation]. + +## Queue Resources + +| Queue | Hostnames | Max Memory | Max CPUS | Max Time | +|--------- |---------------- |------------ |---------- |---------- | +| MSC | compute[01-03] | 32GB | 16 | 336.h | +| Normal | compute[10-29] | 64GB | 16 | 240.h | +| Highmem | compute[04-09] | 128GB | 32 | 2880.h | + +*** + +The configuration profile design is very simple. If your process exceeds 64GB memory or 16 cpus, it is sent to the `highmem` queue. If not, it is sent to the `normal` queue. Please do not use the `MSC` queue, this is reserved for Masters students. + +Take others into consideration when deploying your workflow (do not hog the cluster :pig:). If you need to hammer the cluster with a pipeline, please reach out to me and we can tweak the configuration profile to dispatch jobs to only a handful of compute nodes via hostnames. ## Container Cache -Your workflow containers will be downloaded and stored at `/data/containers` which is accessible to all users. +Your workflow containers are stored under `/data/containers/` which is accessible to all users on lugh.