[SERVER-15354] Implementing cluster index for RocksDB Created: 23/Sep/14 Updated: 16/Oct/15 Resolved: 16/Oct/15 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Storage |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | chensi | Assignee: | Unassigned |
| Resolution: | Duplicate | Votes: | 3 |
| Labels: | rocksdb | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Participants: | |||||||||
| Description |
|
In most cases, querys are focused on the same shardkey, so implementing a cluster index on shardkey will be very friendly for query. RocksDB can sort data by the key, so there can be a way to do this. Thanks |
| Comments |
| Comment by Ramon Fernandez Marina [ 16/Oct/15 ] |
|
This ticket is a duplicate of Regards, |
| Comment by Ben McCann [ 09/Feb/15 ] |
|
Thanks for confirming, Dan. I wasn't entirely sure whether this was referring to a clustered index or not. I guess this is a pretty decent sized feature, but it would be very cool to have. |
| Comment by Daniel Pasette (Inactive) [ 08/Feb/15 ] |
|
This ticket refers to creating a clustered index, that is, storing the data in index-sorted order. See: http://en.wikipedia.org/wiki/Database_index#Clustered. The advantage of having a clustered index on shard key is that logical "chunks" of data which are migrated from shard to shard are also physically stored contiguously on disk. |
| Comment by Ben McCann [ 07/Feb/15 ] |
|
Hi, I'm curious to learn more about this as I'm interested in working on performance improvements to RocksDB. I just got a huge speedup for RocksDB by fixing https://jira.mongodb.org/browse/SERVER-15793 on my machine. What do you mean by a "cluster index on shardkey"? |