Workspaces/Advanced Features/Quotas

From bwHPC Wiki
Jump to navigation Jump to search

Quotas & Limits

Cluster-Specific Workspace Limits

Each cluster has different workspace policies and quotas:

Cluster Default Lifetime Max Lifetime Max Extensions Storage Quota File Limit (Inodes)
bwUniCluster 3.0 1 day 60 days 3 times 40 TiB 20 million files
JUSTUS 2 7 days 90 days Unlimited 20 TiB 5 million files
Helix N/A 30 days 10 times 10 TiB No limit
BinAC2 N/A 30 days 5 times No limit No limit
NEMO2 30 days 100 days 100 times 5 TiB per workspace No limit

Important notes:

  • Default Lifetime: If you don't specify duration, this is used
  • Max Lifetime: Maximum duration you can request
  • Max Extensions: How many times you can extend a workspace
  • Storage Quota: Total space available for your workspaces
  • File Limit: Maximum number of files (inodes) you can create

Note: Limits may change. Check cluster-specific documentation for current values.

Grace Periods for Workspace Recovery

When workspaces expire or are released, they can be recovered for a limited time:

  • Expired workspaces: Recoverable for 30 days (bwUniCluster 3.0/NEMO2), cluster-specific for others
  • Released workspaces: Recoverable until next cleanup cycle (typically during nighttime)

See ws_restore for recovery procedures.

Checking Your Quota Usage

bwUniCluster 3.0, JUSTUS 2

  $ lfs quota -uh $(whoami) /lustre/work
  $ lfs quota -uh $(whoami) /pfs/work9      # bwUniCluster 3.0 specifically

Shows your current usage, quota limit, and available space on Lustre filesystems.

NEMO2

  $ nemoquota                                # Shows all quotas (HOME + workspaces)
  $ df --si $(ws_find workspace_name)        # Check specific workspace

The nemoquota command shows both HOME directory and workspace quotas.

Helix

  $ workquotainfo                            # Shows workspace quota information

BinAC2

  $ df                                       # Check available space

BinAC2 has no quota limits enforced on workspaces.

Managing Quota

If you're running low on quota:

  • Release unused workspaces: ws_release workspace_name
  • Note: Released workspaces count toward quota until cleanup runs (typically during nighttime)
  • Immediate deletion: For instant quota relief, see --delete-data option of ws_release