QA Shift Report Instructions

STAR QA Shift Report

  • If you know that the experiment is taking data, but no fast offline data is appearing in the QA Browser, please make an effort to understand why there is no data (look at the production hypernews, for example) and post this issue to the Electronic Shift Log (under the QA subdetector) if it has not already been posted. Please also remember to close any such opened issues if the problem is resolved.
  • The report is best filled out DURING your shift, not after. This is simply because it is easiest to fill out the form while looking at the data.
  • The plan is: one report per shift, many data entries per report.
  • You should never select the Back button of your web browser during the report process (and you should never need to do so).
  • Do not be afraid to try things out. Until you submit the full report, everything can be modified and reviewed if it has been saved. Also, you can try things out by using a "play" session (see below).
  • If you are experiencing problems, try closing the Shift Report window, clearing your browser's cache and restarting the Shift Report forms - it should not lose your data. Additionally, be sure that cookies and javascript are enabled in your browser. If problems persist, please contact Gene Van Buren at gene@bnl.gov or (631-344-7953).
Choose a session name
You can also continue a session which was started earlier. This name has no meaning other than to identify one QA Shift Report from another in case more than one person is working on a report simultaneously, and to allow one to restore a report that they have already started if their browser or computer crashes. Your chosen session name should appear next to Session in the Menu. You can also make a copy of another session.
If you do not see a session name in the Menu, or the name Unknown appears, please try selecting a session again. Problems will arise otherwise.
If you are learning how to use the forms, you may wish to create a "play" session , where a toy report can be generated, but nothing is permanently recorded (the report is not archived, and issues are not attached to runs nor updated for the "last used" lists). Once a session is created, it cannot be changed to or from being a play session, but such sessions can be copied into new play or non-play sessions.
Fill out the Shift Info form.
Select the Save/Continue button at the bottom of the form when finished. If you are unsure about an item, it can be edited later. If you leave this form (i.e. select any options under Manage Contents) before selecting Save/Continue, you will lose the information you have entered.
Add data entries
Do so for each QA job (histogram set) examined by making the appropriate selection from the Add A Data Entry For... submenu. When combining files for QA, use the pull down menu of trigger types instead of typing in a file sequence. Again, select the Save/Continue button when finished with each data entry. And again, these items can be edited later if necessary, and leaving this form (i.e. selecting anything under Manage Contents) before selecting Save/Continue will result in loss of information on the form.
Focus on issues
Issues are maintained with the Issue Browser/Editor, which can be reached via Open Issue Browser/Editor in the Menu, or from the data entry forms (where only issues for that data type are listed). Issues can then be made active for a given data entry to indicate that the issue was seen there. The intent here is that opened/created issues may be used repeatedly and can be modified or appended later (if more detail is learned, for example) until they are eventually closed/resolved. A name serves as a brief description of each issue, and a detailed description is available via the Browser/Editor, along with any image attachments for the issue. Each issue may also be given tags (by subsystem/category [required], or keywords [optional]) to aid in finding issues or correlating them later. To be listed in the data entry form of a given data type, an issue must be allowed for that type.
You may need to Refresh Issues when working on an entry for any newly created issues to appear. Please make an effort to search through the existing issues before creating a new issue. A filter tool appears to make searching for a related issue easier. 
Manage the contents
After entering each portion of the report, a current contents listing is shown. Each portion may be viewed (by clicking on the item), deleted (by marking the checkbox beside the item and clicking on the Delete button), edited, or copied.
Submit the QA Shift Report
When ready to submit, select Finish & Submit Report. This takes you to the final information to be entered for the report. You can save this information and submit later if you decide you are not yet ready to submit the full report. You can also choose not to delete the session after submitting, so that it can be reused and modified another time.

At any time, you may (using the Menu):

  • Browse or edit issues by selecting Open Issue Browser/Editor in the Menu.
  • View the currently entered portions of the report by selecting View Current Contents.
    You may subsequently edit these files or copy their contents into a new data report.
  • Re-enter or edit the Shift Info by selecting (Re)Enter Shift Info.
  • Erase the entire report and start over by selecting Erase Sesson & Start Over.
  • Stop the current session and choose to start a new one, or continue a different one, by selecting Session.

 

OfflineQA DB documentation

QA Browser

To be written.


QA Shift Reports

Three tables in the OfflineQA database are currently used: QAIssues, QARunFileSeqIndex, and QAShiftReports. Their structure and indexing are defined by these create statements:
CREATE TABLE QAIssues (ID int not null primary key, Name tinytext,
Description text, closed bool default 0,
timeFirst timestamp default CURRENT_TIMESTAMP,
timeLast int default 0, timeFRP int default 0, timeRDP int default 0,
timeRNT int default 0, timeMDP int default 0, timeMNT int default 0,
flagsFRP bool default 0, flagsRDP bool default 0,
flagsRNT bool default 0, flagsMDP bool default 0, flagsMNT bool default 0);
ALTER TABLE QAIssues ADD INDEX(timeFRP,ID);
ALTER TABLE QAIssues ADD INDEX(timeRDP,ID);
ALTER TABLE QAIssues ADD INDEX(timeRNT,ID);
ALTER TABLE QAIssues ADD INDEX(timeMDP,ID);
ALTER TABLE QAIssues ADD INDEX(timeMNT,ID);
CREATE TABLE QArunFileSeqIndex (runYear smallint not null,runDay smallint not null,
run int not null default 0, seq varchar(16) not null, idx tinyint(1) not null,
RepType tinyint(1) not null, RepNum varchar(16) not null, link text,
primary key (runYear,runDay,run,seq,RepType,idx));
ALTER TABLE QArunFileSeqIndex ADD INDEX(RepType,RepNum);
CREATE TABLE QAShiftReports (RepType tinyint(1) not null, RepYear int not null,
RepMonth int not null, RepText mediumtext, RepNum varchar(16) not null,
primary key (RepType,RepNum));
ALTER TABLE QAShiftReports ADD INDEX(RepType, RepYear, RepMonth);
All three tables are given mysql ANALYZE and OPTIMIZE commands each time a new QA Shift Report is submitted (approximately once daily during runs).

It would still be good to add a way to quickly identify in which runs a specific issue was seen.