|EGI Inspire Main page|
|Inspire reports menu:||Home •||SA1 weekly Reports •||SA1 Task QR Reports •||NGI QR Reports •||NGI QR User support Reports|
1. Task Meetings
|Date (dd/mm/yyyy)||Url Indico Agenda||Title||Outcome|
|08/04/2013 to 12/04/2013||http://cf2013.egi.eu/||EGI Community Forum||Accounting Workshop plus contribution to Pay-for-use and Federated Cloud workshops|
2. Main Achievements
- CPU Accounting:
- Regional APEL server code made available on github and documentation for testing sent out to sites on request. South Africa, Germany and Taiwan have requested this so far, feedback from these sites on their progress or any issues has been requested.
- SSM 2.0 and new schema installed in production to receive CAR and EMI 3 APEL Client records.
- Cloud Accounting: the majority of Federated Cloud Resource Providers migrated to publishing cloud accounting data using SSM 2.0 using the new cloud accounting usage record definition. A summary cloud accounting record has been created and data in this format provided to CESGA for work on the cloud accounting portal to progress. The CESGA team have also successfully tested receiving records from the cloud accounting repository using SSM 2.0
- Storage Accounting: additional StAR data received from other test instances, notably the INFN implementation, so that storage records are now received in the test database from 50 sites. It was agreed that we would receive storage data from sites once per day, that all records should include the site information and that the units of data should conform to the StAR. In addition, the StAR and GLUE relationship have been documented.
- Application Accounting: continued work on the integration with APEL, definition of additional usage record fields.
- Billing: meetings with EGI and INFN teams to identify how the billing task may be implemented for accounting data.
- OGF Usage Record work: UR 2.0 document is now in public comments.
3. Issues and Mitigation
|Issue Description||Mitigation Description|
|CPU Accounting: There was a network problem on the 29th March which stopped client sites publishing. This was at the beginning of the Easter holiday weekend and caused several days outage to the APEL consumer system. The result of this was that the server could not accommodate the volume of republished data from the client sites over the coming week, despite additional disk being added.||The service was migrated to a physical server as it had been running on a VM. The physical server is also a higher spec. system with SSDs and this has improved the performance and the system is now running again. (Please note: this is the system receiving data from the old APEL client software versions and that the new system has been designed differently to mitigate these issues. The old system will be retired once client sites have migrated to the new version of the APEL client.)|
|Application Accounting: still does not have a clear use case to develop against.||Requested information/feedback at the EGI community forum.|
4. Plans for the next period
- CPU Accounting: continue work to move other client middleware publishing, e.g. ARC, GridSAFE to production
- Cloud Accounting: send daily summaries of Federated Cloud Resource Providers cloud accounting data to Accounting Portal at CESGA using SSM 2.0.
- Storage Accounting: as more data is received from a greater variety of systems, work with Storage Accounting client teams to evaluate the data received and ensure it is comparable across sites and discuss what form the summaries of storage accounting data will take which can be sent to the portal.
- Application Accounting: formalize the additions needed in the CAR. Investigate other techniques for the application accounting (e.g. fork notifications) as suggested during the EGI Community Forum.
- Billing: assess software currently in use which might be used to provide billing function for accounting data.