This bug is about the runs and reconstruction passes affected by the double remapping of the St1 HV DCS aliases.
LHC12hi
Most of the runs potentially affected (because of a HV sector -the one closest to the beam- of DE202 that is at 1400 V < 1590 V). But the OCDB was manually changed at the time so the PatchSt1DCSAliases was not applied, hence no problem ;-)
| Pass | AliRoot version | approx. reco date | affected |
|---|---|---|---|
| pass2 | v5-05-Rev-30 | apr ‘15 | yes (but OCDB was patched, so no) |
| muon_calo_pass2 | v5-04-Rev-08 | oct ‘13 | no |
LHC13d
Only 1 run affected : 195767, because, again, of DE202 which had a voltage below nominal (1400 instead of 1600 V). Still the efficiency is not zero in this region (see below the recpoints map, built with a special reconstruction pass on SAF3 without any cut on the chamber 2 HV values), but the resolution should be pretty poor.
| Pass | AliRoot version | approx. reco date | affected |
|---|---|---|---|
| pass4 | v5-05-Rev-32-01 | july ‘15 | yes |
| pass2 | v5-03-Rev-28 | jun ‘13 | no |
| mtrg_full | v5-03-Rev-18 | mar ‘13 | no |
| muon_pass2 | v5-03-Rev-18 | mar ‘13 | no |
To confirm the affected versions, tried the recontruction of two chunks of this run, which gives, in the affected pass, a different number of bad pads (to find out, just grep “human” in the rec.log) By looking at the output of the grep command we find two suitable chunks :
/alice/data/2013/LHC13d/000195767//pass4/13000195767000.10/rec.log:According to mask 400be9b (human readable form below) 112848 pads are bad (over a total of 1064008, i.e. 10.61 %)
/alice/data/2013/LHC13d/000195767//pass4/13000195767000.16/rec.log:According to mask 400be9b (human readable form below) 116688 pads are bad (over a total of 1064008, i.e. 10.97 %)
Btw, to find the number of different cases we have :
find /alice/data/2013/LHC13d/000195767/ -type f -name rec.log | xargs grep -i human > tata
cut -d ' ' -f 9 tata | sort -u
112848
116688
(i.e. two different cases for the 1243 reco logs we have for this run)
LHC15g-l
All the pp periods before the bug was fixed are affected, so basically all of LHC15a-l. The first version with the fix is AliRoot v5-07-06 :
git tag --contains e2d1ec0
v5-07-06
v5-07-07
v5-07-08
etc...
The lists of runs potentially affected are below. Potentially affected means there was a non-nominal HV for one of the HV sector (in St1) for which the DCS alias remapping was performed.
List of tested runs for LHC15g
List of tested runs for LHC15h
List of tested runs for LHC15i
List of tested runs for LHC15j