| BNL monitor | Production Clouds DDM PandaMover AutoPilot Sites Analysis Physics data Usage Plots ProdDash DDMDash | |||||
| Update | Not logged in. List users | |||||
| Panda monitor Times are in UTC |
Quick guide to the Panda monitor | |||||
| Wiki Bugs Shift elog CERN elog |
||||||
| Jobs - search Tasks - search Datasets - search Datasets Distribution Sites - see all Applications |
Reporting problems: Report Panda problems of any sort to the Panda Savannah ticketing system.
BNL: Primary production monitor CERN: Monitor for CERN instance of Panda atldq2pro.cern.ch: CERN based backup for BNL primary monitor
Production: Panda Production Operations Dashboard. Summary of Panda production status
Jobs running in Panda: Job links at left list the running, activated (ready for pickup by a pilot), waiting (waiting for input data availability), assigned (brokered and waiting for completion of input data transfer to processing site), defined (awaiting brokerage), finished and failed jobs. Analysis jobs (as opposed to managed production jobs) can be listed separately. The 'old archive' contains all finished/failed jobs older than 3 days. Quick searches: Enter a Panda job name or ID, dataset name or ID, or task name or ID into the appropriate field and hit return in order to do a quick lookup. Summaries: Enter a day count in the desired summary field and hit return to bring up a summary covering the last N days. The 'blocks' summary shows the production datablocks (datasets) currently being processed in the production system, with details on where they are being processed, job states etc. The 'errors' summary shows overall production status at all Panda sites with details of the error conditions encountered. The 'nodes' summary shows worker nodes active at all production sites with statistics on processed jobs and states. Tasks: Task request forms are provided for entering generic, event generation and CTB tasks. The full task list gives statistics on tasks by grid and a listing of all tasks. The task browser allows selection of tasks based on their metadata (physics type, production series, release, output type, etc.) with navigation to datasets associated with the task, where data availability and access information is provided. Datasets: The dataset browser allows browsing of DQ2 datasets based on dataset metadata and site selections. Dataset searches can be done with the search form (with wildcards) or quick search (by name, no wildcards). Listings are available for input datasets (a short list; the datasets which are inputs to the tasks processed by Panda), output datasets (very long -- the datasets produced by Panda production -- but the task browser is more convenient for navigating to produced data), dispatch blocks (Panda internal), and all datasets. Long lists are truncated and useful only to get the overall count and a sample list. Datasets distribution: Dataset replication requests, operational displays from ATLAS DDM Ops for distribution of AODs, RDOs, conditions data, real DAQ data, etc. Subscriptions: Shows DQ2 subscriptions managed by Panda, which handle Panda's data movement. 'Dispatch blocks' are used to dispatch data to processing sites in advance of processing; destination blocks are used to route outputs to archival storage. Sites: Configuration details of sites. Site pages provide access to monitors, jobs running at the site, DQ2 configuration, etc. Logging monitor: Summary of the incident logging sent by Panda components to record Panda activity. Shows job requests (analysis and production) from pilots, by site. Panda system configuration and status: Configuration link, top left, gives system configuration parameters and server status. The system statistics page at left gives overall production summary and stats. |
|||||