[SERVER-9345] Reload sharded (collection) metadata in groups Created: 13/Apr/13 Updated: 10/Dec/14 Resolved: 02/May/14 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Sharding |
| Affects Version/s: | 2.2.4, 2.4.1 |
| Fix Version/s: | None |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | Scott Hernandez (Inactive) | Assignee: | Unassigned |
| Resolution: | Won't Fix | Votes: | 0 |
| Labels: | chunkmanager, configdb, metadata | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Backwards Compatibility: | Fully Compatible |
| Participants: |
| Description |
|
When reloading metadata for a database, all collections or a subset, we could do so in groups and in parallel to reduce the total for the reload. If you have 10000+ sharded collections and need to reload all their metadata in the chunk manager then it can take a long time (esp. if the config server is not local and on a higher latency network). |
| Comments |
| Comment by Greg Studer [ 02/May/14 ] |
|
This is fraught with peril since you need to load metadata in a particular order to correctly talk to shards. Think reducing the metadata load in general and ensuring non-initial loads are fast would solve most of these problems. |