AWS parallelcluster documentation change

Service: parallelcluster · 2025-04-28 · Documentation low

File: parallelcluster/latest/ug/directories-v3.md

Summary

Restructured directory list formatting and added explicit /home directory note

Security assessment

Formatting changes and documentation of default shared directories. While shared directories can have security implications, this change does not introduce new security controls or address vulnerabilities. The /home directory note reflects existing behavior rather than adding security guidance.

Diff

diff --git a/parallelcluster/latest/ug/directories-v3.md b/parallelcluster/latest/ug/directories-v3.md
index ccfea07e5..c231640b0 100644
--- a//parallelcluster/latest/ug/directories-v3.md
+++ b//parallelcluster/latest/ug/directories-v3.md
@@ -9 +9 @@ There are several internal directories that AWS ParallelCluster uses to share da
-`/opt/slurm`
+  * `/opt/slurm`
@@ -11 +11,7 @@ There are several internal directories that AWS ParallelCluster uses to share da
-`/opt/intel`
+  * `/opt/intel`
+
+  * `/opt/parallelcluster/shared (only with compute nodes)`
+
+  * `/opt/parallelcluster/shared_login_nodes (only with login nodes)`
+
+  * `/home (unless specified in SharedStorage)`
@@ -13 +18,0 @@ There are several internal directories that AWS ParallelCluster uses to share da
-`/opt/parallelcluster/shared (only with compute nodes)`
@@ -15 +19,0 @@ There are several internal directories that AWS ParallelCluster uses to share da
-`/opt/parallelcluster/shared_login_nodes (only with login nodes)`
@@ -17 +20,0 @@ There are several internal directories that AWS ParallelCluster uses to share da
-`/home (unless specified in SharedStorage)`
@@ -23 +26 @@ By default these directories are created on the head nodes EBS volume and shared
-When the cluster scales out, NFS exports via the EBS volume may pose performance bottlenecks. Using EFS, you can avoid NFS exports as your cluster scales out and avoid performance bottlenecks associated with them.
+When the cluster scales out, NFS exports via the EBS volume may pose performance bottlenecks. Using EFS, you can avoid NFS exports as your cluster scales out and avoid the performance bottlenecks associated with them.