Loading...

XML

Word

Printable

JSON

Type: Improvement
Resolution: Done
Priority: Major - P3
Fix Version/s: 3.2.12, 3.4.0-rc4
Affects Version/s: None
Component/s: Replication
Labels:
None

Backwards Compatibility:
Fully Compatible
Backport Requested:

v3.2
Sprint:
Repl 2016-10-31, Repl 2016-11-21

Today, replSetStepDown waits for one node to have caught up completely with it. Instead, we can change this behavior to wait for a majority of nodes to catch up completely with it.
In this way, we can avoid all rollbacks that might result from the primary stepping down.

is documented by

DOCS-9449 Docs for SERVER-26747: replSetStepDown should wait for last write to commit

Closed

is related to

SERVER-27339 ReplSetTest shouldn't inspect TestData directly

Closed

SERVER-27117 set writeConcernMajorityJournalDefault to false in nojournal, inmem, and ephemeralForTest variants

Closed

SERVER-27118 Default shutdown command's 'timeoutSecs' argument to 10 seconds

Closed

SERVER-27024 Write additional test for new replSetStepDown behavior

Closed

related to

SERVER-27196 add timeouts to stepdown_needs_electable_secondary.js

Closed

SERVER-27214 Disallow chaining in stepdown_needs_electable_secondary.js and stepdown_needs_majority.js

Closed

SERVER-28110 Set secondaryCatchUpPeriodSecs smaller when requesting remote primary stepdown

Closed

(3 related to)

Assignee:: William Schultz (Inactive)

Reporter:: Eric Milkie

Participants:: Eric Milkie, Githook User, William Schultz

Votes:: 0 Vote for this issue

Watchers:: 8 Start watching this issue

Created:: Oct 24 2016 05:43:06 PM UTC

Updated:: Mar 07 2017 07:04:04 PM UTC

Resolved:: Nov 14 2016 08:49:51 PM UTC

Details

Description

Attachments

Issue Links

Activity

People

Dates