Sharded install causes sync jobs to hang

[trondhindenes]

We tried switching from a single-pod to a sharded one using "dynamicClusterDistribution". After enabling that, jobs would not start, they were just sitting there "waiting to start". No errors in the application-controller logs, it was as if jobs weren't queued up. Anyone else experiencing this? We're having serious problems making argocd react to sync requests etc so I was hoping a sharded setup would help