Wiki TODO

This page lists Anunna documentation that has been published but still needs current or authoritative information filled in. The pages below are usable as they stand, but each contains placeholders — marked in the page source with comments — where someone with current operational knowledge needs to add or confirm details.

To pick up a task: open the page, find the TODO comment(s) in its source, add the information, and remove the item from this list.

Hardware and network

Cluster Architecture Overview — confirm the current interconnect (Omnipath and/or InfiniBand) and topology; add a current architecture diagram (the old schematics were out of date and were removed).
Compute Hardware Overview — fill in the per-node specifications: number of CPU nodes, cores and memory per node, CPU model(s), and whether there are high-memory "fat" nodes; number of NVIDIA GPU nodes, GPUs per node, and host CPU/memory; AMD GPU model and node details.
Network & Security — confirm the interconnect (as above) and document the data-security posture: how data is protected at rest and in transit, access control, handling of confidential or personal data, and backup/retention.
Performance Optimization/Multiple nodes (MPI) — confirm the recommended MPI library and module names, any required srun --mpi=... plugin flag, and whether specific interconnect/fabric tuning is needed. The old mvapich2 + InfiniBand (ib0) example was Breed4Food-era and was removed.

Storage

Quotas — document the Lustre quota model (per-user or per-group, and the default sizes for /lustre/backup, /lustre/nobackup, /lustre/scratch), the command(s) users run to check their current usage, and how to request more space. Only the home-directory quota (200 GB) is currently documented.
Archival Storage — clarify how the manual iRODS/itape workflow relates to Tapeworm, now that Tapeworm manages archival from /archive to tape automatically: is itape still the recommended way to push data to tape, or is iRODS now used mainly for retrieving archived datasets?
Data Lifecycle Policy — add the formal data-retention policy if one exists: how long different data is kept, any maximum retention, and what happens to data when an account or project closes.

Software and workflows

Workflow Engines (Snakemake, Nextflow) — confirm the recommended way to run Snakemake on Anunna (the documented profile is the older Snakemake <8 style; Snakemake 8+ uses the --executor slurm plugin), and write the Nextflow section (loading/installing Nextflow, the SLURM executor configuration, and a minimal example).
Conda for teaching — modernise for current Anunna: the page still describes the old Bright /cm/shared paths and tcl modulefiles. Replace with a Miniforge environment on Lustre and the current course-kernel route (kernel registered at notebook.anunna.wur.nl; see Steps for courses).
Julia — confirm the exact Julia module name(s) and version(s) available in the current buckets.

Support and operations

Maintenance Schedule — confirm where planned-maintenance dates are announced (Apps Portal message of the day, mailing list, Teams) and add the next scheduled downtime, or a link to a live maintenance schedule. Only the general cadence (roughly April/May and October) is documented.
Reporting Usage — document the reports.anunna.wur.nl usage and cost dashboard: what reports it offers, who can see what (your own usage versus your group's), and how to read them.
Managing Group Members — confirm how a PI adds or removes members from their Anunna group (self-service, via the AD group owner, or via the administrators) and document the steps.
Storage Requests — confirm whether storage requests use a dedicated form or the general HPC (Anunna) support service, and any approval or group-owner requirements.
Remote Access (VPN, Gateway) — link the official WUR VPN setup instructions, and document any Anunna-specific gateway or jump-host arrangements.
Resource Allocation Requests — document the resource-allocation process: what can be requested (extra priority/QoS, dedicated nodes, GPU allocations), how it relates to a group commitment, who approves it, and lead times.
External Collaborator Access — document the external-collaborator process: sponsorship and approval, any data/security agreement, account duration and renewal, and access scope.

Policy and governance

Mission and Governance — confirm and document the current governance bodies (for example a steering group and user group) and their membership. The previous roster was from the Breed4Food era and was removed.
Roadmap — add the current roadmap, or link the latest roadmap document. The 2019–2020 version was removed.
Sustainability — add the sustainability / green-HPC policy and any concrete measures: energy sourcing, cooling, hardware lifecycle, and institutional targets.
Policies and Terms of Use — supply the formal terms of use / acceptable-use policy (data ownership and responsibilities, security obligations, consequences of misuse) and confirm the current account-request route and per-group access contacts. The previous access-contact list was from the Breed4Food era and was removed.
Account Application Process — confirm whether WUR users request an Anunna account through the general "HPC (Anunna)" support service or a dedicated account-request form, and what details are needed (group/affiliation, sponsor, intended use).
User's group — document any formal user representation: regular user meetings, how users feed back into decisions, and how that relates to the governance bodies. Currently only the Anunna Users Teams community is documented.
Handling Sensitive Data (GDPR, etc.) — document Anunna's posture on confidential and personal data / GDPR: which data categories may be processed, any data-processing agreements or approvals, protection at rest and in transit, and who to contact for a data-protection assessment.
Grant Support Documentation — provide grant-support material: a standard facility description for funding applications, the acknowledgement text to cite Anunna in publications, cost/quotation support, and who to contact for a letter of support.

Nice to have

FAQ — extend with the questions the HPC team is asked most often (for example quota limits, GPU access, course accounts, data transfer).