Alert.png The wiki is deprecated and due to be decommissioned by the end of September 2022.
The content is being migrated to other supports, new updates will be ignored and lost.
If needed you can get in touch with EGI SDIS team using operations @

EGI-InSPIRE:MS108 apel

From EGIWiki
Jump to navigation Jump to search
EGI Inspire Main page

Notes to contributors

Assessment: (Provide an assessment of the delivery of services over the past year from a managerial perspective; highlight positive areas and areas for improvement; do not include future plans; text should be roughly 1-2 paragraphs)

Score: (assign a numerical score from 1 to 5 with a succinct explanation of what needs to be improved to increase your score – remove numerical description references upon completion) 1 = An unacceptable level of service was delivered

2 = A level of service that was below expectations was delivered

3 = An acceptable service level has been delivered

4 = A level of service that exceeded expectations was delivered, but there is scope for even further improvement

5 = An excellent service has been delivered that should be considered as best practice

Table 4: EGI Global task assessment:Central accounting repository
# Name Assessment Score How to Improve
# Central accounting repository The Central EGI Accounting Repository was kept running this year remarkably well given the laughably small amount of funding (0.25 FTE) allocated to it in SA1. There was one serious incident when a database failure led to a buildup of records waiting to be published which overwhelmed the nodes receiving them. This was resolved by installing a solid state disk for a critical bottleneck in the workflow.

The deployment of the ActiveMQ interface to the repository (developed in JRA1) went smoothly but deployment of the associated client by sites went very slowly. There seems to be no way for EGI to make sites update their middleware, or to read the documentation when they do. The support load due to tickets raising issues covered well by the documentation was much higher than expected.

The R-GMA interface to APEL was finally switched off at the end of February 2011, many months later than planned, due to slowness of sites migrating away from it, as mentioned above.

There is insufficient funding to run the central service. STFC has been heavily subsidising this work. There have been 3+ FTE working on it for most of this year (plus additional sysadmin effort for the hardware used) compared with aggregate funding of just over 1 FTE from EGI and EMI combined. It is not obvious that this level of subsidy can continue in future years.

Management of the team has been straightforward as they are all from one partner but the number of other bodies wanting to interact over accounting has been a challenge.

4 Be more proactive in pushing sites earlier to move to glite-APEL. Run a tutorial at the TF.

Better internal monitoring of the performance of the service and the internal workflows.

More predictable performance. Due varying workload, there is sometimes a large variation in the time between a site publishing and the data being visible in the accounting portal.