Site locality support for file services in a stretched cluster environment

A non-transitory computer-readable medium having instructions stored thereon, which in response to execution by one or more processors, cause the one or more processors of a host to perform or control performance of a method to provide site locality support in a stretched cluster having hosts at different geographic locations, wherein the method comprises:

• determining a site affinity setting of a file server;
• based on the determined site affinity setting and on a workload of a first host at a first site at a first geographic location, locating the file server and corresponding data path components at the first host, such that access by a client to a storage object at a distributed storage system traverses through the first site and avoids cross-site access to a second site;
• in response to a failure at the first host, performing a failover to migrate the file server and corresponding data path components from the failed first host to an active second host;
• and
• in response to restoration of the first host to be active, performing a failback to migrate the file server and at least some of the corresponding data path components from the active second host back to the restored first host, wherein the file server and the at least some of the corresponding data path components are migrated from the active second host to an active third host, instead of performing the failback to migrate from the active second host back to the restored first host, based on a guardrail that specifies an amount of tolerable disruption due to attempting the failback to migrate from the active second host back to the restored first host.

An arbitrator host to provide site locality support in a stretched cluster having hosts at different geographic locations, the arbitrator host comprising:

• one or more processors;
• and
• a non-transitory computer-readable medium coupled to the one or more processors, and having instructions stored thereon, which in response to execution by the one or more processors, cause the one or more processors to perform or control performance of operations that include:
• determine a site affinity setting of a file server;
• based on the determined site affinity setting and on a workload of a first host at a first site at a first geographic location, locate the file server and corresponding data path components at the first host, such that access by a client to a storage object at a distributed storage system traverses through the first site and avoids cross-site access to a second site;
• in response to a failure at the first host, perform a failover to migrate the file server and corresponding data path components from the failed first host to an active second host;
• and
• in response to restoration of the first host to be active, perform a failback to migrate the file server and at least some of the corresponding data path components from the active second host back to the restored first host, wherein the file server and the at least some of the corresponding data path components are migrated from the active second host to an active third host, instead of performing the failback to migrate from the active second host back to the restored first host, based on a guardrail that specifies an amount of tolerable disruption due to attempting the failback to migrate from the active second host back to the restored first host.