Transcript Downlink Segment Status

Downlink Segment Test Status (S15.0)
Russ Laher
August 28, 2006
S15.0 Test Readiness Review, August 28, 2006
Russ Laher (1)
Summary of S15.0 D/L work performed





D/L-group software changes implemented for S15.0 (all critical level 3 or 4)
 56 pipeline-related CRs/AIs (59 for S14.0)
 3 database-related CRs/AIs (12 for S14.0)
 24 ARs (19 for S14.0)
D/L segment-testing to check specific CR/AR/AI changes
 Small cluster of 16 pipeline-drones (mini “ops-like” environment)
 New Solaris-10 operating system
 SODB (fresh S14 “pedigreed” database via Ed Jackson)
 Product archiver run on pipeline results for end-to-end verification
ISTs doing science verification of pipeline products
QA team verification of QA sub-system upgrades
Test plan documents written by the Cog. E.s to help I&T
S15.0 Test Readiness Review, August 28, 2006
Russ Laher (2)
What’s New for S15.0 – IRAC





New C module called “slgst” (CR #7089)

To set Bit 2 (indicates ghost-light corruption) and

To set Bit 3 (indicates stray-light corruption) in Imask image
(see SOSDL-SIS-PD-3030 for Imask-bit definitions)

Requires 4 new calibration files (1 per channel; caltype=“GHOSTMASK”)
Changes to fiducial-frame pipeline (plScriptId=25) (CR #7327)

To query for 2MASS sources

Execute caltrans to fetch calibration images for “slgst” module

Execute the new “slgst” module

New parameters added to w_fiducial_frame.nl
Latency-ensemble pipeline change (plScriptId=23) to get inputs from BCD pipeline
(plScriptId=17), not darkcal pre-processing pipeline ( plScriptId=1) (AR #7378)
New SkyDark selection in BCD pipeline (plScriptId=17) (CR #7301)

Applies to HDR images only, for external repeat >1

Use sky dark from regular (non-HDR) AOR for repeated HDR exposures
New LabDark selection in BCD pipeline (plScriptId=17) (CR #7326)

Applies to channel 4 only

Applies only to exposure times of either 100 or 200 seconds

Applies to HDR images only, for external repeat >1

Use lab dark from regular (non-HDR) AOR for repeated HDR exposures
S15.0 Test Readiness Review, August 28, 2006
Russ Laher (3)
What’s New for S15.0 – IRS








BCD-pipeline change to not propagate uncertainties computed by “droopres”
module (w_irs_droopres.pl) (plScriptId=2015 and many others) (CR #7321)
Modification to the “coadd2d” module to propagate the input uncertainties of
BCDs (plScriptId=2016) (CR #6850)
Added input uncertainty image darkdrift_unc.fits to the “imagest” module’s
command line (plScriptdId=2015 and many others) (CR #7351)
Eliminated “shadow” NaNs made by “imagest” module (mode 1) (AR #7244)
New “snest_irs” module to replace “snestimator” module (w_irs_snestimator.pl)
(plScriptId=2015 and many others) (CR #7010)
New FITS keywords in BCD header for velocity corrections: JULIAN, VX, VY, and
VZ (plScriptId=2015 and many others) (CR #7343)
New non-linearity calibration file used by the “satcor” module (“imagest” mode 2)
(w_irs_satcor.pl) (plScriptdId=2015 and many others) (CR #7352)
Changes to the “extract” module (e.g., flag wavelength bins containing all NaNs
with -9999.0) (several plScriptIds) (CR #7344)
S15.0 Test Readiness Review, August 28, 2006
Russ Laher (4)
What’s New for S15.0 – MIPS


MIPS:Si
 None!
MIPS:Ge
 Change to uprop library and radhit module for Solaris-10 compatibility (Sun
apparently added a new long double version of the <math> library and this
created ambiguity in interpreting whether un-cast scalars are data-typed as
either a double or long double) (executed by BCD pipelines associated with
1500-series plScriptIds) (CR #7336)
 New off-line option for least-squares linear and quadratic fitting of stim
solutions in “interp” module (new option is for offline use for comparing with
results from AZ pipelines only) (executed by BCD pipelines associated with
1600-series plScriptIds) (CR #6477)
 Change to MIPS SED post-BCD processing: Bug fixed in flux computation
(needed to multiply CDELT1 by RESCALE in spectrum module)
(plScriptid=1049) (CR #7255)
S15.0 Test Readiness Review, August 28, 2006
Russ Laher (5)
What’s New for S15.0 – Database



New “releaseLevel” field in all calibration fallback database tables (CR #7165)
 IracFallback
 IrsFallback
 Mips1Fallback
 Mips2Fallback
 Mips3Fallback

Small integer datatype, default=99
Modified database stored function “getCalProductsForRequest” to query on new
“releaseLevel” field in all calibration fallback database tables (CR #7166)

Used by Leopard

No change to interface, only internal queries
New database stored procedures to update new “releaseLevel” field in all calibration
fallback database tables (CR #7184)
 updateIracFallbackReleaseLevel(<fbId>,<releaseLevel>)
 updateIrsFallbackReleaseLevel(<fbId>,<releaseLevel>)
 updateMips1FallbackReleaseLevel(<fbId>,<releaseLevel>)
 updateMips2FallbackReleaseLevel(<fbId>,<releaseLevel>)
 updateMips3FallbackReleaseLevel(<fbId>,<releaseLevel>)

Executable under pipelineOpsRole and pipelineSwRole
S15.0 Test Readiness Review, August 28, 2006
Russ Laher (6)
What’s New for S15.0 – Other






The “angleavg” module was modified to make it robust against < 2 samples
when computing the gyro jitter statistics (AR #7360)
Upgraded “nanlib” library for improved NaN handling on multiple platforms
(e.g., Solaris, Mac, Linux, Windows) (CR #7216)
A reference to “nanvalue.h” has been added to many libraries used by many
modules for porting to non-Solaris platforms (e.g., uprop, dlimage, sortutil, dlcs,
colfit, integrate, …) (CR #7249)
Modified “c2s_trans” and “s2c_trans” modules to allow input columns “delta_...”
to be optional (IRAC bandmerge pipeline; plScriptid=32) (AR #7187)
Bug fix in “bandmerge” module to recognize input SNR for optional statistical
calculations and enlarged output flux fields to handle large fluxes (AR #7180)
Fixed SDAT problem of hanging when displaying an image with all NaNs
(CR #7264)
S15.0 Test Readiness Review, August 28, 2006
Russ Laher (7)
Issues/Concerns

S15 D/L software segment-level testing was conducted under the
new Solaris-10 operating system (previous Solar 8 was used).
 Changes to Makefiles for building D/L software were required
 Comparisons of products generated by Solaris 8 vs. Solaris 10
show no differences
 Perl built under Solaris 8 works find under Solaris 10, so we will
not be re-building Perl for Solaris 10.
S15.0 Test Readiness Review, August 28, 2006
Russ Laher (8)