[SERVER-46118] Make migration_util helpers save the repl set term before finalizing migration Created: 12/Feb/20 Updated: 25/Mar/20 Resolved: 28/Feb/20 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Sharding |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Task | Priority: | Major - P3 |
| Reporter: | Randolph Tan | Assignee: | Randolph Tan |
| Resolution: | Won't Fix | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Sprint: | Sharding 2020-02-24, Sharding 2020-03-09 | ||||||||
| Participants: | |||||||||
| Description |
| Comments |
| Comment by Esha Maharishi (Inactive) [ 25/Mar/20 ] |
|
When I did the write-up above, I missed that the "term check" also checks if the node is primary, meaning the local writes will not keep getting retried if the node is secondary. I amended |
| Comment by Esha Maharishi (Inactive) [ 20/Mar/20 ] |
|
Filed |
| Comment by Esha Maharishi (Inactive) [ 20/Mar/20 ] |
|
Per stakeholder demo, the "retrying indefinitely while secondary" bug is bad enough that we should try to fix it, particularly because a warning message will keep being logged. |
| Comment by Esha Maharishi (Inactive) [ 19/Mar/20 ] |
|
For more context: If the term is saved at the beginning of each operation's loop, the operation must be safe to execute if the term changed just before the term was saved. There are two important cases:
Here are all the operations run in the loop currently, and why they are safe in both cases:
|
| Comment by Esha Maharishi (Inactive) [ 28/Feb/20 ] |
|
Closing as Won't Fix since we will instead do |