On Friday 21st, from 14:00 to 15:00, SSB was inaccessible from outside CERN
On 3rd November, presentation on latest changes and plans of SSB
SAM:
Latest results take more than a minute to load. Still under investigation by DBA and dashboard team. Execution plan has changed, and it is not understood.
Dedicated server (fixes temporary problem changing data)
If there are no problems with this interface, we'll swap the aliases
This week we will remove the old Maintenance column.
Maintenance and BDII. At the moment, if site disappears from BDII, after 48 hours we don't take it into account for the downtime. Shall we increase this to 1 month? Bigger time will also affect entries that have to be deleted
New version of SSB will be deployed in development (only GUI changes, like warnings are presented in different way, detecting 'Flash blockers', or default number of rows')
On 7Th March, with WMAgent_0_7_X release the WMAgent should correctly report information to the Dashboard (job performance numbers will be included later on).
Till now no useful information is reported to the dashboard integration server . (with not useful I mean that all the parameters are reported as 'NotAvailable' value)
New version of the UI will be deployed for ATLAS and LHCb. The CMS development will be done next week (1st week of March, and moving to production 2 weeks later unless anyone complains). This version contains
Nicer looking table for 'Expanded values'
Filtering, paging, sorting by any columns
Faster query
Collectors. Updated last week. Now we can insert ~200 entries per second (during the Christmas break, it took almost 30 min per entry!)
Downtime collector improved with the changes required by Pepe (removing downtimes if they are removed from GOCDB/OSG, and checking if the type of the downntime has changed). These changes will be put in production these week
SSB collectors modified. Instead of 15 independent collectors writing to the DB, they will write to the local file system, and there will be only one inserting things in the database. This has been tested in integration for over a week, and now it will be deployed in production