From af2e034422aacfbe3d83f25295a7a4ef09948f5f Mon Sep 17 00:00:00 2001 From: jamesbeedy Date: Wed, 12 Feb 2025 22:40:42 +0000 Subject: [PATCH 1/4] initialize profiling with influxdb page --- .../enable-job-profiling-with-influxdb.md | 24 +++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 howto/setup/enable-job-profiling-with-influxdb.md diff --git a/howto/setup/enable-job-profiling-with-influxdb.md b/howto/setup/enable-job-profiling-with-influxdb.md new file mode 100644 index 0000000..f519942 --- /dev/null +++ b/howto/setup/enable-job-profiling-with-influxdb.md @@ -0,0 +1,24 @@ +(howto-enable-job-profiling-with-influxdb)= +# How to enable job profiling with InfluxDB + +Charmed-HPC integrates with the [InfluxDB Charm](https://charmhub.io/influxdb) to enable job profiling in slurm. + +This how-to guide explains how enable job profiling by deploying and integrating `influxdb` with charmed-hpc. + +## Prerequisites + +- A [deployed Slurm cluster](#howto-setup-deploy-slurm). + +## Deploy and Integrate InfluxDB + +InfluxDB can be deployed using Juju in a single command. +:::{code-block} shell +$ juju deploy influxdb +::: + +Now [integrate](https://canonical-juju.readthedocs-hosted.com/en/latest/user/reference/juju-cli/list-of-juju-cli-commands/integrate/) the newly deployed influxdb charm with slurmctld. +:::{code-block} shell +$ juju integrate influxdb slurmctld +::: + +Once the InfluxDB charm deployment and integration are complete slurm will be configured to send job profiling metrics to influxdb, enabling the use of the [`sstat`](https://slurm.schedmd.com/sstat.html) command. From f0ac99616cbd86a4b2339095c78e61da4009dd8c Mon Sep 17 00:00:00 2001 From: jamesbeedy Date: Wed, 12 Feb 2025 22:43:19 +0000 Subject: [PATCH 2/4] add refs for job profiling --- howto/index.md | 1 + howto/setup/index.md | 5 +++-- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/howto/index.md b/howto/index.md index 3f16335..fdcc576 100644 --- a/howto/index.md +++ b/howto/index.md @@ -18,6 +18,7 @@ taking you through the setup of your own Charmed HPC cluster. - {ref}`howto-setup-deploy-slurm` - {ref}`howto-setup-deploy-shared-filesystem` +- {ref}`howto-enable-job-profiling-with-influxdb` (howto-manage)= ## Manage diff --git a/howto/setup/index.md b/howto/setup/index.md index f748738..79acc65 100644 --- a/howto/setup/index.md +++ b/howto/setup/index.md @@ -4,7 +4,7 @@ See the how-to guides in this section for the steps to set up and deploy various - {ref}`howto-setup-deploy-slurm` - {ref}`howto-setup-deploy-shared-filesystem` - +- {ref}`howto-enable-job-profiling-with-influxdb` @@ -15,5 +15,6 @@ See the how-to guides in this section for the steps to set up and deploy various Deploy Slurm Deploy a shared filesystem +Enable job profiling with InfluxDB -::: \ No newline at end of file +::: From d47dd3d35340609a20fafaed455f7a6f8301e970 Mon Sep 17 00:00:00 2001 From: jamesbeedy Date: Mon, 24 Feb 2025 22:48:57 +0000 Subject: [PATCH 3/4] add InfluxDB to custom wordlist --- .custom_wordlist.txt | 1 + 1 file changed, 1 insertion(+) diff --git a/.custom_wordlist.txt b/.custom_wordlist.txt index 8e5a343..f5b3b99 100644 --- a/.custom_wordlist.txt +++ b/.custom_wordlist.txt @@ -12,6 +12,7 @@ PKEY RDMA RESource UCX +InfluxDB # Charm/application/service names alertmanager From 2c8aa9d458cb23e38cc7f02667daff0ee2fb5567 Mon Sep 17 00:00:00 2001 From: jamesbeedy Date: Mon, 24 Feb 2025 22:55:07 +0000 Subject: [PATCH 4/4] update custom wordlist --- .custom_wordlist.txt | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/.custom_wordlist.txt b/.custom_wordlist.txt index f5b3b99..cc732bb 100644 --- a/.custom_wordlist.txt +++ b/.custom_wordlist.txt @@ -12,7 +12,6 @@ PKEY RDMA RESource UCX -InfluxDB # Charm/application/service names alertmanager @@ -41,6 +40,8 @@ Terraform terraform Traefik traefik +InfluxDB +influxdb # CLI commands hostname