[SERVER-23733] Remove the mongos chunkSize configuration option Created: 14/Apr/16 Updated: 14/Mar/17 Resolved: 02/Jun/16 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Sharding |
| Affects Version/s: | None |
| Fix Version/s: | 3.3.8 |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | Kaloian Manassiev | Assignee: | Kaloian Manassiev |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||||||||||
| Backwards Compatibility: | Minor Change | ||||||||||||||||
| Sprint: | Sharding 13 (04/22/16), Sharding 14 (05/13/16), Sharding 15 (06/03/16) | ||||||||||||||||
| Participants: | |||||||||||||||||
| Description |
|
Currently, mongos instances can be configured with a chunkSize and autoSplit settings. The first indicates the maximum chunk size after which chunks will be considered jumbo and won't be moved until split and the second indicates whether inserts should automatically split chunks. Since the balancer now will reside on the config server, there is no point in porting these options to mongod as well and instead they should be driven entirely from the values in config.settings. |
| Comments |
| Comment by Githook User [ 02/Jun/16 ] |
|
Author: {u'username': u'kaloianm', u'name': u'Kaloian Manassiev', u'email': u'kaloian.manassiev@mongodb.com'}Message: This change removes the chunkSize startup option from mongos in place of |
| Comment by Githook User [ 04/May/16 ] |
|
Author: {u'username': u'kaloianm', u'name': u'Kaloian Manassiev', u'email': u'kaloian.manassiev@mongodb.com'}Message: Instead, they should pass it as parameter to ShardingTest. This change is in preparation for removing the chunkSize parameter to |
| Comment by Kaloian Manassiev [ 15/Apr/16 ] |
|
renctan, this is correct. However using startup options introduces the possibility of starting different mongos instances with different configurations and opens opportunity for mistake. My thinking is that such balancer behaviour is best controlled from a single source for the entire cluster (or per collection). Also with the plans to eventually move autosplit to the shards, this option will have to go away regardless. |
| Comment by Randolph Tan [ 14/Apr/16 ] |
|
autosplit is not handled by the balancer so shouldn't it be unaffected by this (until we move it to the shards)? |