150 likes | 281 Views
MRO End-to-End Test Status. Ray Arvidson and Keith Bennett November 29, 2006. End-To-End Test Summary. Purpose: Exercise MRO science data flow from instrument team archives to posting of PDS archives
E N D
MRO End-to-End TestStatus Ray Arvidson and Keith Bennett November 29, 2006
End-To-End Test Summary • Purpose: Exercise MRO science data flow from instrument team archives to posting of PDS archives • Approach: Series of delivery tests to identify and correct problems before first delivery to PDS in June of ’07 • Tests involve HiRISE, CRISM, SHARAD, MCS, CTX, and MARCI MRO Instrument Teams • Tests involve Geosciences, Atmospheres, Imaging, and Engineering Nodes
Test Plan – Four Test Phases • Test 1 – May ’06 • Instrument teams provided a single EDR • Transferred to the PDS node via an agreed upon mechanism • Validated EDR against PDS standards and appropriate SISs • Test 2 – July ’06 • EDR Archive with single EDR generated using planned instrument processes • Archive generated as planned (either by instrument team or PDS node) • Delivered via planned delivery mechanism • Validated EDR and Archive against PDS standards and appropriate SISs • Verified availability through PDS catalog search system • Test 3 – Oct ’06 • EDRs - Same as test 2 except with 1 day’s worth of EDRs • RDRs – Same as test 1 with a single RDR • Both EDRs and RDRs assembled into archive volumes • Test 4 – Feb ’07 • Same as test 3 except with 7 days’ worth of EDRs and RDRs • Additional test if needed in ~April/May ‘07
Test Status • Test 1 – May ’06 • Test successfully completed • Final report issued • Several minor discrepancy reports (DRs) issued • All closed in Test 2 • Lien issued – No CTX/MARCI testing because instrument team was not ready • Closed in Test 2 • Test 2 – July ’06 • Test successfully completed • Several minor discrepancy reports issued • All closed in Test 3 • Preliminary report issued
Test 3 (Oct ’06) Status • CRISM/Geosciences • Successfully Completed • Several Minor Discrepancy Reports related to Labels • Some closed, the rest are expected to be closed in test 4 • SHARAD/Geosciences • Successfully Completed • Several Minor Discrepancy Reports related to Labels • Minor issues with labels (expected since RDR SIS still in peer review) • Some closed, the rest are expected to be closed in test 4 • HiRISE/Imaging • Successfully Completed • Minor issues with labels (expected since RDR SIS still in peer review) • MCS/Atmospheres • Successful (not quite finished reviewing all data but no issues seen to date) • CTX/MARCI/Imaging • Status Pending
Errors and Problems EncounteredLabels not meeting PDS Standards or SISs • Example: SOURCE_PRODUCT_ID = { "HRL00002794_00_DF089S_EDR0" = } • Many of these errors were actually in the Instrument teams’ software • Correction 1: instrument team updated software • Correction 2: Sometimes the SIS was changed instead • All errors found in tests 1 and 2 were corrected by test 3 • Only a few of these types of errors were found in test 3 and are expected to be fixed in test 4
Errors and Problems EncounteredLabel References • Label referencing a non-existent file • Example from SHARAD: Line 45 – referenced file: ^PROCESSED_ECHO_TABLE = "R_0083201_001_SS05_700_A.DAT" The actual file is "R_0083201_001_SS05_700_A000.DAT" • Correction 1: instrument team updated software
Errors and Problems EncounteredData Dictionary Errors • Several new keywords or values were missing from Data Dictionary, despite generation of MRO Local Data Dictionary • Some new keywords or values were not entered by the time of the test • Discrepancies between SIS keywords and label keywords • Correction 1: Timely and accurate data dictionary updates reduced errors by test 3 • Correction 2: Updates of SISs or automatic label generation software by Instrument teams reduced keyword errors • Highlights current problems with managing local data dictionaries!
Errors and Problems EncounteredLVTool Errors • Label Validation Tool (LVTool) sometimes failed to correctly handle valid keywords/values, although none prevented LVTool from running to completion. • Example: “LVTool reports an error on a BIT_COLUMN when an ITEMS field is included. LVTOOL indicates a BITS field is needed but the BITS field is options when there is an ITEMS field (as per PDS Standards).” • Correction 1: Some errors corrected with updated version of LVTool • Correction 2: Some have been deferred to the new Validation Tool
Data Transfer and Access Results • No data transfer / access errors • Data Transfer/Access Examples • CRISM – Test 3 - ~7GB via Data Brick • SHARAD – Test 2 – ~3GB via FTP from Italy • HiRISE – Test 3 - ~8GB validated via remote access to Data Node • CTX/MARCI – Test 2 – 300MB • MCS – Test 3 – 630MB via FTP
PDS Data Flow and Web Access Results • Updated PDS catalog with MRO test data • Demonstrated PDS web access to MRO catalog data and science data at the nodes • No major problems
MRO Test Summary • Next Step - Test 4 – February ’07 • Objectives: • Exercise data transfer • Exercise data validation • Reduce potential first-delivery problems • Useful for pushing instrument teams to finish SISs
Future PDS End-2-End Plans • Missions planning to do E2E testing: • Phoenix • MESSENGER • LRO • MSL
Key Lessons for the Future • E2E reduces first-delivery problems • Test timing is crucial • Too early and products/SISs not ready • Too late and they interfere with operations • Test goals need to be clear: • Test products (production, validation, etc.) • Test handling (delivery, archive assembly, publication) • Number of tests depends on test goals and product/instrument complexity
PDS Lessons and Questions for Future E2E Tests • Improve Inter-PDS test communication • MRO E2E inter-PDS communication was often slow and relied too much on the mission • Suggest having PDS-Only E2E telecons outside mission archive working groups • Discussion Topics: • Does PDS or the mission drive the tests? • Who is the customer of the test results? • What is the role of PDS test coordinator? • Should there be a common definition of what constitutes an Discrepancy Report?