You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Ivan Bessonov (Jira)" <ji...@apache.org> on 2022/12/19 12:12:00 UTC

[jira] [Created] (IGNITE-18429) ItTablesApiTest#testTableAlreadyCreatedFromLaggedNode is flaky

Ivan Bessonov created IGNITE-18429:
--------------------------------------

             Summary: ItTablesApiTest#testTableAlreadyCreatedFromLaggedNode is flaky
                 Key: IGNITE-18429
                 URL: https://issues.apache.org/jira/browse/IGNITE-18429
             Project: Ignite
          Issue Type: Improvement
            Reporter: Ivan Bessonov


No link here, failure can be reproduced locally.

Here's the flow of events:
 * MetaStorageManager#storeEntries starts its execution, revision is equal to 2, for example;
 * At the same time, MetaStorageManager#updateAggregatedWatch is called by the inhibitor in the test;
 * "storeEntries" reads an applied revision value, it's 1.
 * "updateAggregatedWatch" reads an applied revision value, it's still 1.
 * New watch is started, using revision 2 as a base.
 * "storeEntries" writes data into the vault, applied revision becomes 2.
 * New event from new watch appears:
 ** revision of the event is 2, because that's the base;
 ** revision read from Vault is also 2;
 ** we get an exception that these two cannot be equal.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)