In order to run any parametric test, one should check for the assumptions. few important assumptions which you should check on your data before carrying principal component analysis :

1: You should have sampling adequacy, which simply means that for PCA to produce a reliable result, large enough sample sizes are required. The method to check this assumption, Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy is used. if you get KMO value close to 1, that means this assumption is met, however values above .5 or .6 are acceptable.

2: your data should follow normal distribution. Kolmogorov–Smirnov test is used to check the normality of data set. if your dataset does not follow normal distribution, which is quite obvious for real time data, you can use transformation menhods such as 

log transformation, square root transformation, inverse etc.

3: There should be no significant outliers. 

4: there should be linear relationship between all variables.

Using PCA you don't have to check normality the KMO and Barttelt test of adequacy are enogh, KMO must be greater than 0.5 and as Dziuban & shrkey (1974) mentioned that if

KMO have valeu 0.9's then the sample is marvelous for the analysis

 MO have valeu 0.8's then the sample is meritorious

KMO have valeu 0.7's then the sample is middling

KMO have valeu 0.6's then the sample is medicore

KMO have valeu 0.5's then the sample is miserable

good luck

Thank you Nida R for valid suggestions