User guide

gShare lets many people share a Kubernetes GPU cluster — in fractional or exclusive slices — and accounts for usage in credits. This guide walks through the user console: the screens you use to run sessions, keep data, and manage credits.

• Signing in — first login, language, theme, the cluster selector.
• Dashboard — balance, burn rate, allocation, availability.
• Sessions — what a session is, its states, the list screen.
  • Creating a session — the five-step wizard, field by field.
  • Connecting — VS Code, JupyterLab, the web terminal.
  • Managing a session — pause, resume, restart, live usage.
  • Waiting in the queue — why a session waits and what to do.
  • Ending a session — terminating, billing, what survives.
• Data and volumes — persistent storage: scopes, types, the list.
  • Creating a volume — scope, type, access mode, size.
  • Using a volume in a session — mounting, paths, who has it open.
  • Sharing a volume — granting, revoking, leaving a share.
  • Resizing, locking, deleting — quota, locks, safe deletion.
• Wallet and credits — how billing works: hold, consume, settle.
  • Transaction history — every movement in the ledger.
  • Requesting credits — asking your group, and tracking it.
• Account — profile and membership.
  • Password — changing it, first sign-in, resets.
  • My limits — the resource policy, and asking for more.
  • Notifications — the bell and its history.

Administrative features — organizations, groups, users, credits, clusters, monitoring — are in the Administrator guide.

Only GPU time costs credits

CPU sessions and volumes are free. Your balance never falls while no GPU session is running.