[DOCS-615] Documentation on localThreshold internals Created: 16/Oct/12 Updated: 29/Nov/12 Resolved: 30/Oct/12 |
|
| Status: | Closed |
| Project: | Documentation |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | Andre de Frere | Assignee: | Sam Kleinman (Inactive) |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||
| Participants: | |||||
| Days since reply: | 11 years, 16 weeks, 1 day ago | ||||
| Description |
|
Given the level of documentation on --localThreshold it would be a common expectation that setting this option will cause an immediate switch of queries falling below the threshold. However, as the value is calculated on a moving average, and only done once every 10 seconds, it can take some time for the steady state ping to emerge. This means some queries will hit secondaries below the threshold until the steady state is reached. Some documentation on how we reach the threshold calculated value would clear up any confusion. |
| Comments |
| Comment by auto [ 30/Oct/12 ] |
|
Author: {u'date': u'2012-10-30T11:55:53-07:00', u'name': u'Sam Kleinman', u'email': u'samk@10gen.com'}Message: |
| Comment by Sam Kleinman (Inactive) [ 16/Oct/12 ] |
|
Changing the setting requires a restart of the mongos, which should reset the cached average, and provide a more expected behavior. I can add a note to the documentation that says the mongos calculates latency using a rolling average of response times, and as a result it may take a little while for actual changes in ping times to affect observed behavior. Having said that, this seems like the kind of thing that might be worth thinking about as a behavior change to the C++ driver (and therefore the mongos) if it's the kind of thing that is unintuitive and actually causing problems. |