driveAtlas: fix large1 bay numbering, row-major vs column-major mismatch

The grid renderer placed bays column-major (1-5 down the left column,
6-10 middle, 11-15 right) while every wiring comment in the file was
written row-major (A1=1, B1=2, C1=3, A2=4, ...). This silently mislabeled
almost every populated bay on screen -- e.g. bay 7 rendered at B2 when the
wiring docs meant A3.

Also corrects the phy7 (Slim-SAS #1 Port 1) wiring note: previously
flagged as an unverified guess with no drive present. osd.14 was moved
onto this port after its onboard SATA port started corrupting the OSD
under load; physical cable-color tracing of the middle stack (silver/
silver/orange/silver bottom-up) confirms the original B Slot 2 (bay 5)
guess was correct.

Verified live against large1: fixed script test-run over SSH shows
bay 5 (sdi/osd.14) and bay 7 (sdd/boot SSD) both drawn in the correct
cells, matching the wiring documentation for the first time.

driveAtlas.sh

@@ -835,7 +835,8 @@ generate_nuc_layout() {
 # Rack ears with D-handles at both ends; slim I/O strip on the right.
 #
 # Bay numbering is a driveAtlas LOCAL convention (the chassis documents none):
-#   1-5 = LEFT module top->bottom, 6-10 = MIDDLE, 11-15 = RIGHT.
+#   row-major across modules: A1=1, B1=2, C1=3, A2=4, B2=5, C2=6, ... C5=15.
+#   (Matches the wiring comments on SERVER_MAPPINGS["large1"] below.)
 #
 # Args:
 #   $1 - Hostname to display in the layout header
@@ -951,9 +952,9 @@ generate_large1_layout() {
     local eartop='┌───┐' earhole='│ ▄ │' earcapT='│╭─╮│' earbody='││ ││'
     local earcapB='│╰─╯│' earbot='└───┘'
     local labs l1 l2 l3
-    printf -v l1 '%*s%-*s' 5 '' 13 'LEFT 1-5'
-    printf -v l2 '%*s%-*s' 3 '' 15 'MIDDLE 6-10'
-    printf -v l3 '%*s%-*s' 3 '' 15 'RIGHT 11-15'
+    printf -v l1 '%*s%-*s' 5 '' 13 'LEFT A'
+    printf -v l2 '%*s%-*s' 3 '' 15 'MIDDLE B'
+    printf -v l3 '%*s%-*s' 3 '' 15 'RIGHT C'
     labs="${l1} ${l2} ${l3}"
     local d grid E
 
@@ -966,8 +967,11 @@ generate_large1_layout() {
             0)  grid="$gtop" ;;
             10) grid="$gbot" ;;
             1|3|5|7|9)
+                # row-major: bay N = A/B/C module, slot ceil(N/3) -- matches
+                # the SERVER_MAPPINGS wiring comments (A1=1,B1=2,C1=3,A2=4,...)
                 i=$(( (d + 1) / 2 ))
-                grid="│${cells[$i]}│ │${cells[$(( i + 5 ))]}│ │${cells[$(( i + 10 ))]}│" ;;
+                local base=$(( (i - 1) * 3 ))
+                grid="│${cells[$(( base + 1 ))]}│ │${cells[$(( base + 2 ))]}│ │${cells[$(( base + 3 ))]}│" ;;
             *)  grid="$gmid" ;;
         esac
         E="${DA_HL}${earbody}${DA_SR}"
@@ -995,8 +999,8 @@ generate_large1_layout() {
 
     printf '%s\n' "  ${DA_ST}LEGEND    ${DA_OK}▌${DA_ST} drive present      ${DA_DIM}░${DA_ST} bay empty$(_l1_rep ' ' 50)${DA_RST}  "
     printf '%s\n' "  ${DA_ST}NOT IN GRID   m2-1 ${DA_HL}${m1p}${DA_SR}   m2-2 ${DA_HL}${m2p}${DA_SR}   internal ${DA_HL}${exp}${DA_RST}  "
-    _l1_txt 'BAYS      numbering is a driveAtlas LOCAL convention (the chassis documents none): 1-5' "${DA_ST}"
-    _l1_txt '          LEFT module top->bottom, 6-10 MIDDLE, 11-15 RIGHT -- identify a drive by SERIAL.' "${DA_ST}"
+    _l1_txt 'BAYS      numbering is a driveAtlas LOCAL convention (the chassis documents none):' "${DA_ST}"
+    _l1_txt '          row-major A1=1,B1=2,C1=3,A2=4,...,C5=15 -- identify a drive by SERIAL.' "${DA_ST}"
     _l1_txt 'NOTE      no backplane, no per-bay LEDs, not hot-swap: every drive is cabled direct to the' "${DA_DIM}"
     _l1_txt '          board/HBA, and is serviced by removing the TOP COVER, not through the front door.' "${DA_DIM}"
 
@@ -1083,7 +1087,12 @@ declare -A SERVER_MAPPINGS=(
     #     Port 1 (phy3) -> C Slot 1 (bay 3)    Port 2 (phy2) -> unconnected
     #     Port 3 (phy1) -> A Slot 1 (bay 1)    Port 4 (phy0) -> A Slot 2 (bay 4)
     #   Slim-SAS #1 (SILVER cables, connector furthest from bracket):
-    #     Port 1 (phy7) -> B Slot 2 (bay 5)  -- cable present, no drive yet
+    #     Port 1 (phy7) -> B Slot 2 (bay 5) -- confirmed 2026-08-09: osd.14
+    #       (28TB OOS28000G, serial 0002QQ2N) moved onto this port from a
+    #       failing onboard SATA port. Verified by cross-checking the middle
+    #       stack's cable colours bottom-up (silver/silver/orange/silver =
+    #       B5 stuck/B4/B3 ata-7/B2 this port), which matches this mapping
+    #       exactly -- the original "no drive yet" guess was correct.
     #     Port 2 (phy6) -> B Slot 5 (bay 14) -- STUCK, cannot remove
     #     Port 3 (phy5) -> B Slot 4 (bay 11)
     #     Port 4 (phy4) -> A Slot 3 (bay 7)