You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Roman Puchkovskiy (Jira)" <ji...@apache.org> on 2023/04/06 11:21:00 UTC
[jira] [Resolved] (IGNITE-18170) Deadlock in TableManager#updateAssignmentInternal()
[ https://issues.apache.org/jira/browse/IGNITE-18170?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Roman Puchkovskiy resolved IGNITE-18170.
----------------------------------------
Resolution: Fixed
Fixed in IGNITE-18203
> Deadlock in TableManager#updateAssignmentInternal()
> ---------------------------------------------------
>
> Key: IGNITE-18170
> URL: https://issues.apache.org/jira/browse/IGNITE-18170
> Project: Ignite
> Issue Type: Bug
> Reporter: Roman Puchkovskiy
> Assignee: Roman Puchkovskiy
> Priority: Major
> Labels: ignite-3
> Fix For: 3.0.0-beta2
>
> Attachments: threads_report.txt
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Currently, {{TableManager#updateAssignmentsInternal}} is fully synchronous. The scenario is as follows:
> # {{updateAssignmentsInternal}} starts a RAFT group for a partition
> # {{FSMCallerImpl}} finds out that its applied index is below the group committed index, so it starts to apply the missing log entries in its {{init()}} method (this is still done synchronously)
> # While doing so, it invokes {{{}PartitionListener{}}}, which tries to execute an insert
> # To make an insert, a PK is needed, so it the insertion code tries to obtain a PK from its future like this: {{pkFuture.join()}}
> # That future is completed from {{{}IndexManager#createIndexLocally(){}}}, which is invoked by {{ConfigurationNotifier}} later than {{updateassignmentsInternal}} in the same thread
> # As a result, the PK future cannot be completed before the sync {{updateAssignmentsInternal}} finishes its job and returns, and it cannot finish its job before the PK future is completed
> We should make {{updateAssignmentsInternal}} async.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)