When there are several initial-loads runs simultaneously, then when the first of them will finish then it is going to pick up the first batch of instance ids without considering the current request-id which will lead to retrieving mixed instance ids which can relate to different request ids(harvesters) or duplicated ids by the reason that it can retrieve the same instance ids which hold different request ids, such duplicates lead to an error.
- Need to consider the request-id within instances list query.
- Need to consider the request-id when deleting instances via resumption token request. Previously resumption token from one request has caused the cleaning of all other instances which match instance ids processed by such resumption token. Such behavior was not correct, only instances related to request id within resumption token as well should be deleted
- Records should be normally returned via resumption token flow at any time even when there are several requests done already.