I have posted a couple of pages that outline the steps I took to fix the bad fills for the 200 GeV transverse running period. These served as an excellent refresher for me as I went through the process of getting the v124 files ready to upload. I would now like to outline the good and bad fills, and the runs/fills lost in one concise post.

Attached is a document, badList.original.txt, which outlines the fills that were flagged as "bad" by spbitSeeker.tcl, and those that needed QA. All-in-all there are 19 fills that need help. The first step was to remove those fills which had two or less runs, as these will not be fixable since for most of them the error says there's a difference between the starting fill number and the cdev record fill number. So once we remove these fills we have removed the following runs (you can see these outlined with more information in testSummary.original.csv attached below):

This leaves 13 fills that need to be QA'ed. To start I start with those fills that have no error message. This means that begEndComp.tcl found differences in the beginning and ending v124 files that are created by spbitSeeker.tcl. To do this for most of the runs I check the ending time of the fill (that is found in GMT on the online RLB) to see if CDEV gives the output we expect at that time (fill number, polarization, etc.) if not I search for a time where all of these match up and reset the ending time of the fill. For those fills where no runs were removed but the ending time had to be reset, I will outline them below and give the ending time of the fill that will be uploaded to the spin database:

F16436 - Reset ending time to 11:09 GMT on 2/17/2012 instead of 11:17:06 GMT 2/17/2012

F16484 - Reset ending time to 2012-02-29 06:41:41 GMT instead of RLB ending time of 2012-02-29 06:51:42 GMT

F16509 - Reset ending time to 2012-03-05 15:03:42 GMT instead of RLB ending time of 2012-03-05 15:13:42 GMT

F16541 - Reset ending time to 2012-03-12 00:59:58 GMT instead of RLB ending time of 2012-03-12 01:09:58 GMT

The rest of the 9 bad fills, two of them need no QA. F16462 initially had a P4 pattern which switched to a P1 pattern in the middle of a run in the fill, which was very strange. After looking through it several times we emailed Dmitri and he informed us that the P4 pattern was the real pattern and that the switch was just a change of the pattern name not a real change in which physics was affected, so we upload this fill with a P4 pattern. F16543 was what we believe to be a glitch in the code since it was the final fill in the list. This had no problems at the ending time, or with the beginning fill number even though it had a *fb error, so it's uploaded as-is.

The final 7 fills are those which runs had to be removed to make the fill usable. Some had differences in the v124 files, but some had *fb errors (meaning the beginning of the fill number doesn't match the cdev record for the beginning of the fill) thus we had to remove some runs at the beginning of the fill depending on when the fill number and pattern became consistent with the rest of the fill:

So in the end we had to remove a total of 6 fulls fills (for a total of 7 runs) and modify 7 other fills' runs (a total loss of 10 runs) for a total loss of 17 runs overall. The final documents are added below (goodList.final.txt and testSummary.final.txt) and these have been uploaded to the spin database.