Run9 data production (p+p 200GeV) was made with the DbV20100301.

On March 15, 2010 Dmitry filled tpcAnodeHVavg DB table for Run-9 data. The common bielive is that the table entries uploaded by Dmitry on March 15 are superier compared to ones used during the data production as these entries more realistically represent actual state of the anode HV system during the data taking.  I am not sure which table (tpcAnodeHVavg or tpcAnodeHV) and what entry date timestamps were used during the data production.  I have checked with Yuri that data production used tpcAnodeHVavg as well. 

Xianglei found that for some sectors/padrows anode HV voltage values (this is not real anode HV, but calculated effective HV) are different when running embedding with DbV20100301 and DbV20110301 (with this stamp entries uploaded by Dmitry get used) .

When producing the real data while anode voltage values in the DB are wrong (lets say HV for some channels is ON while actulaly it was OFF) it is in general not a big deal, since there were no hits in the detector. 

Since the HV information is used to decide whether some part of the detector is alive or dead,  having a wrong HV information (ON instead of OFF) in simulation become a problem - some channels can be treated as active unnecessary.

The proposed solution is:

Create copy of the tpcAnodeHVavg table entries inserted into the DB on March 15, 2010 with an entry time set such, that this copy will be picked up when one uses the DbV20100301.