You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@bookkeeper.apache.org by "Ivan Kelly (JIRA)" <ji...@apache.org> on 2012/11/21 14:37:58 UTC
[jira] [Updated] (BOOKKEEPER-475)
BookieRecoveryTest#testSyncBookieRecoveryToRandomBookiesCheckForDupes()
iterates too much
[ https://issues.apache.org/jira/browse/BOOKKEEPER-475?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Ivan Kelly updated BOOKKEEPER-475:
----------------------------------
Attachment: 0001-Reduce-amount-of-iteratons-for-testSyncBookieRecover.patch
> BookieRecoveryTest#testSyncBookieRecoveryToRandomBookiesCheckForDupes() iterates too much
> -----------------------------------------------------------------------------------------
>
> Key: BOOKKEEPER-475
> URL: https://issues.apache.org/jira/browse/BOOKKEEPER-475
> Project: Bookkeeper
> Issue Type: Bug
> Reporter: Ivan Kelly
> Assignee: Ivan Kelly
> Fix For: 4.2.0
>
> Attachments: 0001-Reduce-amount-of-iteratons-for-testSyncBookieRecover.patch
>
>
> This test exists to detect the case where a bookie is taken down, and the same another bookie is selected to replace it, but the other bookie is already in the ensemble. BOOKKEEPER-106 fixed this, so it shouldn't happen. However we still need to check. We iterate 10 times, but this is a parametered test, so effectively we iterate 40 times. There's 3 possible bookies to select, two of which are already in the ensemble, so a single iteration, if selecting randomly has a 66% chance of selecting a bad bookie. We have to run four times anyhow because it's a parameterized test, giving us a 66*4 % chance of hitting the bug if we only have one iteration.
> Therefore, we should only iterate once. Also, we don't need to write much data. We only care about whats in zk for this test really.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira