Can error be completely eliminated?

In an experiment to find acceleration due to gravity $(g)$ using a simple pendulum,the time period of $0.5 \ s$ is measured from the time of $100$ oscillations with a watch of $1 \ s$ resolution. If the measured value of length is $10 \ cm$ known to $1 \ mm$ accuracy,the accuracy in the determination of $g$ is found to be $x \%$. The value of $x$ is:

The maximum error in the measurement of mass and length is $4 \%$ and $3 \%$ respectively. The error in the measurement of density of a cube will be (in $\%$)

The maximum error in the measurement of resistance,current,and time for which current flows in an electrical circuit are $1 \%$,$2 \%$,and $3 \%$ respectively. The maximum percentage error in the calculation of the dissipated heat will be (in $\%$)

$A$ metal wire has mass $(0.4 \pm 0.002) \, g$,radius $(0.3 \pm 0.001) \, mm$ and length $(5 \pm 0.02) \, cm$. The maximum possible percentage error in the measurement of density will nearly be $.......\%$

$A$ physical quantity $Q$ is found to depend on observables $x, y$ and $z$,obeying the relation $Q = \frac{x^3 y^2}{z}$. The percentage errors in the measurements of $x, y$ and $z$ are $1\%, 2\%$ and $4\%$ respectively. What is the percentage error in the quantity $Q$ (in $\%$)?