TelCal Release Notes
TelCal 1.5.41 (08 Sep 2011)
1.5.41 was deployed to production today (9:00 am).
- Upgraded bdfReader from 1.0.2 to 1.0.3 and evla-commons from 1.6.4 to 1.6.6.
- TelCal - JIRA - MC-14 - support for reference antennas and subbands to be used by TelCal in TelCal and MCAF
- Added reference antenna to the PNT file.
- Honoring reference antenna and subband preferences received from MCAF.
1.5.39 was deployed to production & test today @ 15:52.
- evla-telcal-1.5.39 (10aug2011)
- Stopped allocating direct system memory in the BdfReader; instead all memory is held within the heap now.
- Changed the max heap size from 2048MB to 3500MB.
- Changed the initial heap size from 512MB to 3500MB.
- This will correct the OutOfMemoryErrors we've been seeing on observations with large BDF files.
1.5.38 was deployed to test and production (2011-07-12, 02:47pm).
- evla-telcal-1.5.38 (12jul2011)
- Pointing used to start on subscan 3. Now it starts on subscan 2.
1.5.36 was deployed to test and production (2011-04-07, 11:09am).
- evla-telcal-1.5.36 (07apr2011) * Updated the BdfReader to parse BasebandName enum String using fromValue(String) rather than valueOf(String) to fix the 3-bit parse issue. * Updated evla-bdfReader dependency to 1.0.1. * Updated evla-sdm dependency to 1.2.5. * Updated evla-telcal version to 1.5.36.
TelCal 1.5.35 was deployed to production (2011-02-07, 2:00pm) and test (2011-02-07, 1:50pm).
- evla-telcal-1.5.35 (07feb2011)
- Fixed NPE writing Bandwidth to the PNTDTL file.
TelCal 1.5.34 was deployed to production (2011-01-28, 9:44am) and test (2011-01-28, 8:45am).
- evla-telcal-1.5.34 (28jan2011)
- Added bandwidth to the PNTDTL file.
- Increased the precision of the Elevation and Azimuth columns in the PNTDTL file.
- Stopped printing the antsol solution to the log file. It's too big and not useful.
TelCal 1.5.33 was deployed to production (2011-01-24, 2:58pm) and test (2011-01-24, 2:10pm).
- evla-telcal-1.5.33 (24jan2011)
- Added Elevation and Azimuth into the PNTDTL file so that Corey's FluxGen program can create the CAL_POINTING table.
TelCal 1.5.32 was deployed to production (2011-01-19, 3:00pm) and test (2011-01-19, 2:48pm).
- evla-telcal-1.5.32 (19jan2011)
- Updated MJD formatted time to a plain MJD time in the PNTDTL file.
- Removed array-config-based pad name from the PNT and DEL files. This additional naming of the pad is unneeded and broken.
1.5.31 was deployed to production (2011-01-11, 2:45pm) and test (2011-01-07, 2:17pm).
- Since 1.5.30 never went into production, here is the list of changes in both .30 and .31 that are now in production.
- evla-telcal-1.5.30 (05jan2011)
- Changed the log file max size from 100MB to 10MB. It still keeps up to 10 rolling log files.
- Added some more detailed log messages.
- Updated SDM dependency to version 1.2.3.
- Enabled the console logger and got rid of System.out.println calls.
- Now checks to see if the SdmMessageProcessor thread being updated (with a new SdmInfoMessage) is still alive. If it is not, it removes it from the list and creates a new one to replace it.
- evla-telcal-1.5.31 (07jan2011)
- Catching, logging, then rethrowing Errors in each TelCal thread.
- Updated logging by adjusting console logging settings and remove more System.out.println calls.
- Turned off the ReadAheadThread in the BdfReader call.
- It was sometimes leaving these thread open, consuming memory and leading to OutOfMemoryErrors.
- Before we use it again, the BdfReader's ReadAheadThread needs to be debugged.
- Updated install file to capture stderror as well as stdout from the command line.
TelCal 1.5.28 was deployed to test and production. (2010-11-01 9:30 AM MT)
- evla-telcal-1.5.28 (01nov2010)
- Antsol's initial guess is no longer dependent on data from the first antenna (e.g. ea01) but instead uses data from the first non-flagged antenna.
- Only set the flagged and zeroed flags if they are set. This will reduce the multicast size by not sending false flags over and over again.
TelCal 1.5.27 was deployed to test and production. (2010-10-04 1:20 PM MT)
- evla-telcal-1.5.27 (04oct2010)
- Adding two flags flags to antsol to represent two error conditions:
- flagged - means the data was flagged in the SDM or flagged internally in Antsol for having more than five "bad spectrums" (very small amplitudes)
- zeroed - Antsol didn't process this data because it had zero "weight", which is what I expect when the CBE spits out all zeros. (The zeroed flag fixes an error where the previous hint was being treated like a valid solution.)
- Handles blank or bad bdfLocation attributes that don't point to the BDF file.
- SdmSocketListener now logs the SdmInfo message. This will help debug when SdmInfoMessageProcessor isn't doing what we expect.
- Commented out zero checks in calculateDelay and calculatePhase since the flags (flagged & zeroed) now cover this.
- Updated pom file to use evla-commons 1.5.7, which contains the latest antsol schema with the new flags: flagged & zeroed.
TelCal 1.5.25 was deployed to test and production. (2010-09-01 12:21 PM MT)
- evla-telcal-1.5.26 (01sep2010)
- Updated project name in POM file.
- Updated jaxb plugin in POM file to not use verbose logging.
- TelCal now runs Antsol for the scanIntent CALIBRATE_BANDPASS.
- Added sourceEncoding of UTF-8 to the POM file.
TelCal 1.5.25 was deployed to test and production. (2010-08-18 11:46 MT)
- evla-telcal-1.5.25 (18aug2010)
- Now filling the optional mjdTime (double) field in the antsol document.
- Now internally using AntsolType.mjdTime instead of AntsolType.iat.
- This was changed in the AntsolTextFileWriter, Delay, Focus, and Pointing.
- AntsolMessageSender was updated to include mjdTime in its shallowCopy method.
- Updated TelCalStandAlone to work with prettified XML in the SdmInfoMessages.tmp file.
- Put the PNTDTL file in the production shared directory (/home/mchost2/evladata/telcal/) since folks want to see it to examine the pointing at different subbands.
TelCal 1.5.24 was deployed to test and production. (2010-08-16 10:37 MT)
- evla-telcal-1.5.24 (16aug2010)
- Before only 1 subband was used for pointing, the second one (or the first if there was only one). Now all subbands doing pointing independently and all the results are logged in the PNTDTL file. The ptgsol file (and corresponding multicast) and the PNT file continue to only contain results for the 1 subband, as before. Pending analysis of the data collected in the PNTDTL file this process will be updated.
- Removed the TelCalConstants.PTGSOL_VALIDATE_DATA constant. It is no longer needed.
- Now AntennaData only creates PositionData objects for subbands that exist, rather than the max of 16 subbands every time.
- Added numIntegrations to AverageStats output.
TelCal 1.5.23 was deployed to test (2010-08-10 8:41 MT) and production (2010-08-10 9:46 MT).
- evla-telcal-1.5.23 (10aug2010)
- Modified the PeekTextFileWriter so that there is always a space delimiting the offsets, beamWidths and amplitude.
- Data files produced by TelCal (GN, PNT, PNTDTL, ptgsol, FOC, DEL, stats) now use the entire datasetID as the name of the file and no longer prepend the day-of-the-month to it.
- State field (AKA subMode) completely removed from Pointing.
- Antsol secondary hints is now done by baseband-polarization rather than by one or the other. This will prevent hints from allowing solutions to fall through to other IFs.
- Removed unneeded subscanIntent check from Pointing. It was leftover from the old, 10-subscan pointing method.
- Updated install.sh script to build TelCal by extracting the requested version of TelCal from the Maven repository.
TelCal 1.5.22 was deployed to test (2010-08-04 14:54 MT) and production (2010-08-05 1322 MT).
It contains changes needed for the new pointing script.
- evla-telcal-1.5.22 (04aug2010)
- General clean-up of code. No functional changes.
- Commented out the state==0 kludge for pointing.
- Updated eval-commons dependency to version 1.5.4.
TelCal 1.5.21 was deployed to test and production. 2010-07-30 16:10 MT
- evla-telcal-1.5.21 (30jul2010)
- Changed the good estimated beam width range in Focus to [2.0,4.0].
- Changed out file names from datasetID up to the first period to datasetID down to the second from last period. This handles the case of an observation with multiple execution blocks which have identical lup-to-the-first-decimal datasetIDs.
TelCal 1.5.20 was deployed to test and production. 2010-07-28 12:23 MT
- evla-telcal-1.5.20 (28jul2010)
- Removed redundant logging.
- The runTelcal.sh now includes a redirect for standard out to go to output.log.
- Updated the focus solver to use log amplitude.
- Added some comments describing future changes that will be needed when a Focus table is available.
TelCal 1.5.19 was deployed to test and production. 2010-07-26 15:53 MT
- evla-telcal-1.5.19 (26jul2010)
- Updated dependencies to evla-commons 1.5.3 and evla-sdm 1.1.5.
- Updated repositories in the POM.
- Updated expected lustre mount point from the CBE.
- Fixed a bug in the secondary hint in antsol when data was flagged by changing the hints to be tracked by baseband instead of by polarization.
- In Pointing, Az and El are now set to the ON_SOURCE position's Az and El rather than the first available position's Az and El. Also, each new pointingSolution within a scan uses its own ON_SOURCE Az and El rather than just using the Az and El from the previous solution.
TelCal 1.5.18 was deployed to test and production. 2010-07-21 13:38 MT
- evla-telcal-1.5.18 (21jul2010)
- In the GN file Source was changed from 10 to 16 characters and moved to the last column.
- When running in standAlone mode, the stats and PNTDTL file are written to the root log dir like the other output files.
- Log the full antsol message as XML. Corey will need this data for his switch power calculations.
- For Pointing and Focus skip the first two subscans in a scan. There is typically low quality data near the scan boundary.
- Updated assembly plugin in the pom to use the attached goal so that the jar-with-dependencies gets installed and deployed to the maven repositories.
TelCal 1.5.17 was deployed to test and production. 2010-07-14 13:43 MT
- evla-telcal-1.5.17 (14jul2010)
- Fixed pointing by taking the Ln of the amplitudes before calculating pointing. This results in more accurate pointing offsets. The Estimated Beam Width calculation changed in turn which results in more consistent beam width estimates.
- Pointing: After testing we determined the good range for estimated beam width should be [0.85, 1.25]. After more data is obtained we may tighten this up by 0.05 on each side.
TelCal 1.5.16 was deployed to test and production 2010-07-12 15:18 MT
- evla-telcal-1.5.16 (12jul2010)
- Added creation of the stats averages script when installing TelCal.
- Optional state field is now really optional for Pointing.
- Removed left over, now defunct, TelCalAdminConsole components.
- Added more details to performance stats collection, specifically for
- BDF File access times.
- Added MJD as the first column in the GN file.
- Update to use the new BdfReader read-ahead functionality.
- Name the TextFileWriter files based on the start date of the observation rather than on the creation date of the text file.
- 14 Jul 2010