Release Notes - DataServer - Version 1.3.0

Introduced better interface for infile and branch file lists. Fixed couple of minor bugs.

Release Notes - DataServer - Version 1.2.5

Updated root v4 release v1r0p7 directory, for both "classic" and "DC2" versions of the server.
Added DataServer CVS module tag labels to page footers.

Release Notes - DataServer - Version 1.2.4

This version enables user to define dataset files that will be pruned or peeled. The files are supplied by either InfileList File (a file containing paths to these root files), on Infilelist (which is cut/pasted from somewhere else). This release supports ROOT 5 release v2r0p5 of pruner and peeler scripts.

Fixed a bug where a pruner task from history once resubmited was registered in the history as a peeler task (although it was a pruner task). [DATASERV-28]
Fixed a bug where some of the runnames starting with 0 had been ignored when an event list was supplied (e.g., if there was a run 025000 and user would supply run 25000 in the eventlist, this run was not taken into account. This was an old piprline bug where runnames where recorded as strings and had runname entries in the run table like 0100 or 0100-2000. Note that none of the runnames of type 0xxx-xxxx are related to tasks that this DataServer supports currently.
Fixed a bug where user was able to supply a peeler job without event list or event file, causing the runPeel.pl to abort. Now the user is sent to the error page.
Fixed a bug where the user event list was growing during session, each time the user would supply a new event list, during the single session. Quite nasty.
Moved max file size option out of expert options.
Gave a shot on improving UI interface look and feel.

Release Notes - DataServer - Version 1.2.3

This version of the DataServer package is specifically for ROOT v4 users and is addressing various problems that have come up recently, specifically:

Updated to new release of peeler/pruner directory for v4 root files (version v1r0p4). Problems fixed:
1. pointing to /afs release directories that had disappeared
2. running on real (NRL) data rather than MC data
3. 1. large run/event numbers (affecting the code release extraction mechanism)
  2. lack of "lib" directory in non-opt EM releases starting with v6
  3. NRL data trees have corrupt TTreeIndexes and must be rebuilt
  4. Oracle entries for some datatypes have repeated filenames
One workaround (3.d.) requires the user to supply $PEEL_INFILELIST so this option must be available to peel certain datatypes. There should be no change to Pruner operation. In the front end (this DataServer), the PEEL_INFILELIST is computed from the event list (runnames), so user does not need to supply PEEL_INFILELIST as it is automatically suppliled to the batch job. Supplying user infilelist actually is ignored.
Fixed a bug with empty spaces in infile lists
Fixed a bug with debug selection
Removed hints for default values for expert options
Fixed a bug when timestamps in database are NULL, which caused the finding of last batchId to fail.

Release Notes - DataServer - Version 1.2.2

We introduced support for root v4 and root v5 provided files. All the jobs that are in current table of tasknames are requiring root v4 files, except those that have tasknames with "v9" string in it v9, or are Beam-test011 and above. If there is a new taskname in the table, it will be assumed to require root v5 files. However, as an expert, user can override the pruner and peeler executable directory by specifying PRUNE_DIR and PEEL_DIR to point to whatever directory it pleases, assuming that batch job has access to that directory for execution of runPruner.pl and runPeeler.pl. User setting overrides any other settings. Be careful when supplying PRUNE_INFILELIST to use the right version of root files and if necessary override the default root version v5 with the correct version.

Fixed Bugs

[DATASERV-20] - Pruner errors
[DATASERV-23] - Pruner/Peeler should use -G glastgrp

Improvements

Introduced singleton Controller, that hands out database connections and logger to other objects. Introduced EnvVariables object that is subclassed to PrunerEnvVariables and PeelerEnvVariables, each setting its set of environment variables needed by perl executables. The allowed constants are strongly typed using Java enum.

Release Notes - DataServer - Version 1.2.1 and earlier

Fixed Bugs (release 1.2.1)

[DATASERV-22] - DataServer is estimating 0 time for pruner jobs.

Fixed Bugs (release 1.2.0 and earlier)

[DATASERV-13] - Trouble accessing backgndDC2-GR-v7r3p5
[DATASERV-14] - If the user does not choose max file size, the value passed to the batch job is bogus
[DATASERV-15] - The PEEL_DIGI env var does not appear to be set properly
[DATASERV-22] - DataServer is estimating 0 time for pruner jobs.

Improvements

This release has switched to new user interface for the pruner and peeler, and the old interface phased out. Major changes are done for support of expert user options, header files and tab based navigation. Internally most of the code is converted to JSP 2.0 API, although not in its entirety. This should make code maintenance easier.

[DATASERV-16] - More information in the email from the pruner

New Features

[DATASERV-8] - Should have "expert options" on peeler/pruner
[DATASERV-10] - Users should be able to see which version of the dataserver they are using
[DATASERV-18] - It would be nice to have tabs for navigation among pages

Back To Welcome Page