Garden rose production involves field culture for rootstock production and cultivation of grafted plants. When roses are replanted at the same site, reduced growth, stunting and discoloration of roots are observed. This phenomenon is well-known as rose replant disease (RRD), which results in considerable economic losses. Until now, the causes of RRD are largely unknown and measures to overcome RRD are lacking. In contrast to apple, research on replant disease in roses is scarce. In this study, we have carried out a greenhouse pot experiment, using either untreated or gamma irradiated RRD soils from two sites. Slow release fertilizer was added to the soils before planting seedlings of the rootstock Rosa corymbifera ‘Laxa’. After eight weeks, shoot and root growth were recorded, roots were examined by detailed microscopic analyses and secondary metabolites were analyzed by gas chromatography-mass spectrometry. Rhizosphere samples were taken and their respective microbial communities were analyzed by amplicon sequencing of the 16S rRNA gene for bacteria and archaea as well as the ITS2 and cox2 region for fungi and oomycetes, respectively. Finally, segments of surface-disinfected roots were placed on 523 medium and outgrowing endophytic bacteria were isolated and identified. For both soils, significantly higher shoot and root biomass were observed for plants growing in irradiated compared to untreated soils. Roots were darkened, deformed and finally damaged in their outer cell layers when growing in untreated RRD soil. This corresponded to higher concentrations of two catechin derivatives and gallic acid in roots in this RRD soil compared to roots in the irradiated variants. Twenty-six endophytic bacterial isolates were obtained from roots that were affiliated to 15 different bacterial genera. Rhizosphere microbial community compositions not only differed significantly between soils of the two sites, but also between treatments (untreated versus gamma irradiated) for bacteria, fungi and oomycetes. The identification of genera differing in relative abundance in the different soils and treatments will provide a deeper insight in causal agents of RRD as well as antagonists or beneficials. Future analyses should include different rootstock species in order to identify RRD tolerant germplasm.