SGAS maintenance project
SGAS is an accounting system for Grid and HPC systems. Initially rooted within the Swedish Grid project SweGrid, it is currently used to record resource usage within the national infrastructures coordinated by SNIC and UNINETT Sigma and for the Nordic participation in the WLCG collaboration. SGAS is also spread within the Baltic countries and Switzerland. SGAS is a complete suite with LUTS as the server part and BART as the LRMS (Local Resource Management System) reporter part.
SGAS is broadly in use in the Nordic countries within e.g. the Nordic Tier-1, Sweden (SweGrid, SAMS [SNIC Accounting And Metrics System]), Norway and Finland and preferred over competitive solutions like e.g. DGAS (Distributed Grid Accounting System). So therefore NeIC, SNIC and UNINETT Sigma AS have entered a collaboration agreement to continue to fund the maintenance and development of SGAS.
More information about the SGAS suite and links to the newest releases suite can be found at https://sgas.github.io/
Project Organization
Head of steering group: Michaela Barth
Head of reference group: Erik Edelmann
Project Lead: Erik Edelmann
Project Details
Roadmap
Milestone 1
Released on January 22nd, 2014
Common
- (M1) Adding/Managing Bug trac.
- (M1) Plugin store (collecting also not supported, community provided plugins)
- (M1) Setup "new" source code repository (GitHub?)
Bugs
- (M1) Problem with long running, large jobs
- (M1) UTF-multi byte chars (are currently rejected).
Milestone 2
Released on June 10th, 2014
Improvements
- (M2) Improve error reporting from SGAS
- (M2) Update current storage implementation from StAR draft to 1.2 final.
- (M2) Refactor parts of the code for easier maintenance
- BART
- (M2) Client side verification (Add verification step into bart-reporter & storage-reporter)
Extension
- (M2) nagios plugins
Milestone 3
Planned release in January 2015
Improvements
- (M3) get a plug-in infrastructure so add-ons can be added
- (M3) Add SAMS custom query interface as a plugin.
- (M3) make idtimestamp default!, might be already done
- (M3) add GOLD support for Bart (contributed from .is/.no)
- Cloud accounting (UR-2)
- (M3) what can the cloud stacks use? --> check
- (M3) Documentation review
- remove wrong documentation!
Extension
- (M3) Command for remove records for a cluster in a safe way
- Host scale factors.
- (M3)-> remove it from the core and make plugin instead.
Milestone 4
Planned release in May 2015
Improvements
- (M4)Extend the reporting views -> if it takes time, it can take time, more important to get this done the right way
- separate module(s)?
- removing existing views???
- (M4) Cloud accounting (UR-2)
- Depending on the outcome of M3
Milestone X
Planned release in fall 2015
Improvements
- Reporter
- tool for verify star records.
- Change BART archive directory
- Lots of files problem (BART uses one directory per usage record)
- Support of LRMS? Slurm, Moab, Gold
- support for SLURM and Torque already included, but problems with newer versions of LRMSs
- Not supported -> should be community provided plugin.
Extension
- (M3 -> no rather SAMS contribution) Migrate SAMS custom query interface into mainstream SGAS
- Migrate NeIC APEL reporting into mainstream SGAS
- SSM2?
- contribution might be best choice, rather specific and not of generic use
- Host scale factors.
- Put WLCG views into separate module??
Project Assessment
Weekly meetings
- Weekly status meetings on Friday 10:00 via XMPP
Quarterly reports
- Quarterly Report 13Q4
- Quarterly Report 14Q1
- Quarterly Report 14Q2
- Quarterly Report 14Q3
- Quarterly Report 14Q4