[
https://issues.apache.org/jira/browse/IGNITE-12808?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17130755#comment-17130755
]
Ivan Daschinskiy commented on IGNITE-12808:
-------------------------------------------
>>> DynamicEnableIndexingBasicSelfTest - I'm not sure it is really test
>>> anything as grid is pre-started once, so the order of cache creation and
>>> index enabling doesn't matters in all tests except the first one.
I destroyed and created cache before each test. Why it doesn't test anything? I
don't understand you
>>> ynamicEnableIndexingConcurrentSelfTest - cache is started on client, then
>>> indexing is enabled on same client. It doesn't covers case if index
>>> information propagated to client correctly.
No, indexing is enabled not on client, but on srv1. I added stop of srv1 after
enabling and all works (if cache is replicated, of course, if partitioned --
data loss occurs")
{code:java}
@Test
public void testClientReconnect2() throws Exception {
Assume.assumeTrue(cacheMode == CacheMode.REPLICATED);
// Start servers.
IgniteEx srv1 = ignitionStart(serverConfiguration(1));
ignitionStart(serverConfiguration(2));
// Start client.
IgniteEx cli = ignitionStart(clientConfiguration(5));
srv1.cluster().state(ClusterState.ACTIVE);
createCache(srv1);
loadData(srv1, 0, NUM_ENTRIES);
// Reconnect client and enable indexing before client connects.
IgniteClientReconnectAbstractTest.reconnectClientNode(log, cli, srv1,
() -> {
try {
enableIndexing(srv1).get();
Ignition.stop(srv1.name(), true);
}
catch (IgniteCheckedException e) {
throw new IgniteException("Failed to enable indexing", e);
}
});
awaitPartitionMapExchange();
assertEquals(NUM_ENTRIES, query(cli, SELECT_ALL_QUERY).size());
performQueryingIntegrityCheck(cli);
}
{code}
Test above passed and everything is ok
> Allow create tables for existing caches
> ---------------------------------------
>
> Key: IGNITE-12808
> URL: https://issues.apache.org/jira/browse/IGNITE-12808
> Project: Ignite
> Issue Type: New Feature
> Components: sql
> Reporter: Mikhail Cherkasov
> Assignee: Ivan Daschinskiy
> Priority: Major
> Time Spent: 1h
> Remaining Estimate: 0h
>
> If you have a big cache with a lot of data and you need to index it, right
> now you have to destroy cache and create a new one to index your data. Or
> create a new cache with a table and reload it to data to the new cache which
> definitely is time-consuming and super inconvenient.
> I believe we can allow users to create tables for existing caches.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
