Skip to content. | Skip to navigation

Personal tools
You are here: GES DISC Home Additional Features Technology Lab S4PM 5.7.1 RELEASE NOTES

S4PM 5.7.1 RELEASE NOTES

Release Date

August 5, 2005

Functional Changes

  1. Suspend/Resume Bug Fixed - A bug in the suspend/resume capability via the S4PM Job Monitor window was fixed. The 'Resume' button had been disabled after selecting 'Suspend'. This should be fixed now.

  2. Support for Input from Datapool - This release now supports the setting up of symbolic links in the INPUT disk pool to data residing on a local disk somewhere outside of S4PM. It is intended for On-Demand processing where, with ECS, data are distributed via FTP pull Synergy 4 path. Rather than copying the data over from Datapool to the S4PM INPUT disk pool, symbolic links are instead set up. This can only be used if the data reside on a disk visible to the machine where S4PM is running (i.e. mounted). To invoke this feature, set the parameter $input_symlink_root in the Stringmaker/Stringmaster string configuration file. It also requires that the parameter $s4pm_email be set in the Stringmaker/Stringmaster global configuration file. Note ECS users that you must also configure all input data types for Synergy 4 path.

  3. On-Demand Performance Enhancements - A number of changes have been made to address performance in On-Demand processing. The sort order in the Request Data and Find Data stations have been changed to 'FIFO' (first in, first out). In addition, the Find Data station in On-Demand is configured to recycle work orders every 5 minutes rather than sitting on them for the time specified in the algorithm configuration files. This means that in On-Demand, the timers specified in those algorithm configuration files are ignored. At the end of 5 minutes, if data has not been found, the work order is recycled and goes to the back of the queue. During the 5 minutes a job is running, the manual overrides such as IGNORE_OPTIONAL still function normally. Because of this configuration, jobs in Find Data will never fail due to missing data.

  4. File Accumulation Production Rule Fix - Fixed bug in Select Data that was affecting algorithms using the file accumulation production rule and the pre-processing offset option.

  5. Prepare Run Bug Fix - Fixed a logic bug in the Prepare Run station that caused two outputs of the same data type and data time in the generated PCF to have the same exact file name (under some circumstances). This resulted in all but one of the files getting clobbered.

  6. File Name Extension Bug Fix - Fixed a small cosmetic bug in the Register Data station that caused data ordered via the Composed Data Request tool to have the file name extension .hdf.hdf.

  7. Production History Export Bug Fix - Fixed a bug in Run Algorithm that prevented Production History (PH) files from being exported to the ECS archive (TT 15441).

  8. Stub File Deletion Fix - Fixed bug that prevented Request Data stub files from getting properly cleaned out (TT 15490).

Detailed File Changes

PAN.pm

  • Modified a reference to PDR to S4P::PDR.

  • Updated write() to support optional FTP-Push of PAN (backward compatible).

s4pm_derived.cfg

  • Modified so that in On-Demand, the $cfg_sort_jobs is set to 'FIFO'. This has proven to increase performance.

  • Reconfigured Find Data for On-Demand to work with s4p_repeat_work_order.pl.

s4pm_find_data.pl

  • Added a new option to support running the script with s4p_repeat_work_order.pl in On-Demand strings. The new option is -recycle and its argument is the number of seconds to wait before recycling the work order.

s4pm_global.cfg

  • Added new parameter: $s4pm_email. It is required when $input_symlink_root is set, but otherwise ignored.

s4pm_prepare_run.pl

  • Fix for problem that showed up in AIRS processing of their summary browse product. The problem was that for two different LUNs, the output file names were identical resulting in one getting clobbered. The code should have checked for file name uniqueness, but the logic didn't work. The logic was fixed.

s4pm_register_data.pl

  • Removed a line that was adding a .hdf file name extension back into the file name formed. This is no longer needed and was causing a double .hdf extension to be added, e.g.: .hdf.hdf.

s4pm_request_data.pl

  • Fixed bug whereby request data stub files were being incorrectly named and therefore, not getting cleaned out when proxy data types were in use, typically in On-Demand strings (TT 15490).

s4pm_run_algorithm.pl

  • Fix a problem having to do with Production History files whereby the data type (PH) was not being set correctly in the PDR for moving the PH files into the archive (TT 15441).

s4pm_select_data.pl

  • Fix in get_process_start_time(). It was missing a line to apply a pre-processing offset if the file accumulation production rule was used. This missing line was added back in.

s4pm_split_services.pl

  • Modified to consider a delete from archive flag of 'G' to be the same as 'N'.

s4pm_string.cfg

  • Added example and description of new parameter: $input_symlink_root. This parameter, intended for data ordered via FTP pull, causes data in the INPUT disk pool to be symlinked to data in the $input_symlink_root.

s4pm_stringmaker.pl

  • A missing close statement in update_configurator() was causing the s4pm_configurator.cfg file to be written as a zero length file. A close statement was added to fix this.

  • Modified so that it would only add the FTPPASSWORD parameter to the ACQParmfile if the parameter $input_symlink_root isn't set.

s4pm_stringmaker_derived.cfg

  • Modified so that in On-Demand, the $cfg_sort_jobs is set to 'FIFO'. This has proven to increase performance.

s4pm_stringmaker_global.cfg

  • Added new parameter: $s4pm_email. It is required when $input_symlink_root is set, but otherwise ignored.

s4pm_stringmaker_string.cfg

  • Added example and description of new parameter: $input_symlink_root. This parameter, intended for data ordered via FTP pull, causes data in the INPUT disk pool to be symlinked to data in the $input_symlink_root.

  • Reconfigured Find Data for On-Demand to work with s4p_repeat_work_order.pl.

s4pm_stringmaster.pl

  • Modified so that it would only add the FTPPASSWORD parameter to the ACQParmfile if the parameter $input_symlink_root isn't set.

s4p_subscribe.pl

  • Modified references to S4P::S4P::Subscription::* to S4P::Subscription::*.

stationmaster.pl

  • When job fails due to non-log files being present, it will now print out their names in the station.log file.

Subscribe.pm

  • Updated documentation (corrected %cfg_data_url with %cfg_url in POD to reflect what is in the code).

Tkjob.pm

  • Fixed the suspend/resume functionality in configure_job_buttons().

 

Document Actions
NASA Logo - nasa.gov
NASA Privacy Policy and Important Notices
Last updated: Dec 08, 2010 11:06 AM ET
Top