This simulation I did years ago shows why a snubber diode is needed. In the sim, there are three identical inductive loads being switched on-off by voltage-controlled switches. The red trace V(pulse) is the control signal that turns on all three switches. The switches are on when the red trace is positive +1v) , and off when it is negative -1V.
The bright green trace V(nodiode) is the inductive kick out of the unsuppressed coil. The high voltage developed as the coil turns off burns the points of the switch, and couples to nearby circuits, both capacitively, and inductively.
V(diode) lt, blue and V(hiside) dark blue shows the voltage with both a low-side switch and hi-side switch when the diodes are added. Note that with the low-side switching, the voltage across the opening switch is clamped one diode drop above 12V, while with high-side switching, it is clamped one-diode drop below ground.
Finally, the bottom plot panel is for Mike Odom, to show that voltage across the diode V(diode)-V(12v) dark green trace ranges from -12V (same as the supply voltage) to +1V (while the diode is forward biased by the decaying current). Note the forward current through the diode I(d1) pink trace starts at 3A same as the coil current I(L2) gray trace, but it decays rapidly to zero. The nominal 1A diode will have no trouble with an infrequent 3A short duration pulse...
Back to Suraj, your diode outlined in red has to be there to suppress rel2 (to prevent noise and to protect the contacts of Rel1). However, you still didn't answer the question about how you are suppressing the REAL inductive load, the one not shown on your schematic???