US 11,816,029 B2
Adjustment of garbage collection parameters in a storage system
Alyssa Proulx, Boulder, CO (US); and Wei Sun, Boulder, CO (US)
Assigned to NetApp, Inc., San Jose, CA (US)
Filed by NetApp, Inc., San Jose, CA (US)
Filed on Mar. 10, 2022, as Appl. No. 17/691,588.
Application 17/691,588 is a continuation of application No. 16/856,228, filed on Apr. 23, 2020, granted, now 11,288,186.
Prior Publication US 2022/0197789 A1, Jun. 23, 2022
Int. Cl. G06F 12/02 (2006.01); G06F 12/0864 (2016.01); G06F 11/30 (2006.01)
CPC G06F 12/0253 (2013.01) [G06F 11/3006 (2013.01); G06F 11/3037 (2013.01); G06F 12/0246 (2013.01); G06F 12/0864 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A method comprising:
managing, by each of a plurality of metadata services of a distributed storage system, a set of block identifiers (IDs) for data blocks of a volume hosted by a respective storage node of a plurality of storage nodes of the distributed storage system;
maintaining, for each metadata service of the plurality of metadata services, a probabilistic filter containing information indicative of those of the data blocks that are in use;
monitoring an efficiency level of a garbage collection process performed by the distributed storage system by calculating and combining a plurality of false-positive rates (FPRs), wherein a given FPR of the plurality of FPRs represents an estimated FPR of the probabilistic filter of a given metadata service of the plurality of metadata services that is independently determined based on a set of parameters of the probabilistic filter, and wherein for a particular round of the garbage collection process during which the distributed storage system is in a particular state, the efficiency level of the garbage collection process is calculated based upon an efficiency set determined for the volume; and
responsive to determining the efficiency level is below an efficiency threshold:
updating the probabilistic filter of a particular metadata service of the plurality of metadata services to increase the efficiency level by adjusting at least one parameter of the set of parameters of the probabilistic filter of the particular metadata service; and
performing a subsequent round of the garbage collection process on the data blocks of the volume managed by the particular metadata service based on the updated probabilistic filter.