US 11,841,734 B1
Synchronous block level replication across availability zones
Andrey Arkharov, Kirkland, WA (US); Andrei Burago, Kirkland, WA (US); Jonathan Forbes, Bellevue, WA (US); Anton Sukhanov, Bellevue, WA (US); and Fabricio Voznika, Kenmore, WA (US)
Assigned to Google LLC, Mountain View, CA (US)
Filed by Google LLC, Mountain View, CA (US)
Filed on Dec. 15, 2021, as Appl. No. 17/551,914.
Application 17/551,914 is a continuation of application No. 15/893,262, filed on Feb. 9, 2018, granted, now 11,232,131.
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 16/10 (2019.01); G06F 16/27 (2019.01); G06F 9/455 (2018.01); G06F 16/178 (2019.01); G06F 16/20 (2019.01)
CPC G06F 16/275 (2019.01) [G06F 9/45558 (2013.01); G06F 16/178 (2019.01); G06F 2009/45583 (2013.01); G06F 2009/45595 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A method for maintaining block level replication in a distributed storage system, comprising:
maintaining, in one or more memories in a first zone of the distributed storage system, a first replica including a first copy of a disk;
maintaining, in the one or more memories in a second zone of the distributed storage system, a second replica including a second copy of the disk;
writing, by the disk, in a first mode when both the first replica and the second replica are healthy, wherein in the first mode a successful operation must be acknowledged by both replicas;
detecting, with one or more processors, that the first replica is unhealthy;
writing, by the disk, in a second mode when the first replica is unhealthy, wherein in the second mode a successful operation must be acknowledged by the second replica; and
creating, by a virtual machine attached to the disk, a replication lease that indicates at least one of which replica is trusted or which replica is unhealthy, wherein the lease has an expiration time based on a severity of degradation of the first replica that is detected to be unhealthy.