[
https://issues.apache.org/jira/browse/IGNITE-16726?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Denis Chudov updated IGNITE-16726:
----------------------------------
Description:
Scenarios:
1.
- start node A, then start node B
- create table Table1
- stop node B
- check that that exception is thrown on attempt of usage of Table1, as
majority is lost
- start node B
- check that the table Table1 is recovered in TableManager on node B
2.
- start node A, then start node B
- create table Table1 with 1 partition existing on node A
- restart node B
- check that the client to table Table1 is recovered in TableManager on node B
3.
- start node A, then start node B
- create table Table1
- create table Table2
- stop both nodes
- start both nodes with different configuration (e.g. network port changed)
- check that the Table1 and Table2 are recovered on both nodes
4.
- start node A, then start node B
- create table Table1
- stop node B
- make configuration updates, their count should be greater than metadata
difference that is acceptable for recovery to finish (e.g. 100 tables with
alter and drop of some of them)
- start node B
- check that all configuration updates are applied correctly on node B
5. same as 4, but node A should also be stopped while node B is being
recovered, to check that the recovery on the node B continues when the node A
starts again and continue providing metastorage updates.
was:
Scenarios:
1.
- start node A, then start node B
- create table Table1
- restart node B
- check that the table Table1 is recovered in TableManager on node B
2.
- start node A, then start node B
- create table Table1 with 1 partition existing on node A
- restart node B
- check that the client to table Table1 is recovered in TableManager on node B
3.
- start node A, then start node B
- create table Table1
- create table Table2
- stop both nodes
- start both nodes with different configuration (e.g. network port changed)
- check that the Table1 and Table2 are recovered on both nodes
4.
- start node A, then start node B
- create table Table1
- stop node B
- make configuration updates, their count should be greater than metadata
difference that is acceptable for recovery to finish (e.g. 100 tables with
alter and drop of some of them)
- start node B
- check that all configuration updates are applied correctly on node B
5. same as 4, but metastorage should be compacted before the node B starts
again.
> Add new tests checking recovery functionality
> ---------------------------------------------
>
> Key: IGNITE-16726
> URL: https://issues.apache.org/jira/browse/IGNITE-16726
> Project: Ignite
> Issue Type: Task
> Reporter: Denis Chudov
> Assignee: Denis Chudov
> Priority: Major
> Labels: ignite-3
>
> Scenarios:
> 1.
> - start node A, then start node B
> - create table Table1
> - stop node B
> - check that that exception is thrown on attempt of usage of Table1, as
> majority is lost
> - start node B
> - check that the table Table1 is recovered in TableManager on node B
> 2.
> - start node A, then start node B
> - create table Table1 with 1 partition existing on node A
> - restart node B
> - check that the client to table Table1 is recovered in TableManager on node B
> 3.
> - start node A, then start node B
> - create table Table1
> - create table Table2
> - stop both nodes
> - start both nodes with different configuration (e.g. network port changed)
> - check that the Table1 and Table2 are recovered on both nodes
> 4.
> - start node A, then start node B
> - create table Table1
> - stop node B
> - make configuration updates, their count should be greater than metadata
> difference that is acceptable for recovery to finish (e.g. 100 tables with
> alter and drop of some of them)
> - start node B
> - check that all configuration updates are applied correctly on node B
> 5. same as 4, but node A should also be stopped while node B is being
> recovered, to check that the recovery on the node B continues when the node A
> starts again and continue providing metastorage updates.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
