flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yun Tang (Jira)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-10954) Hardlink from files of previous local stored state might cross devices
Date Wed, 03 Jun 2020 12:31:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-10954?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17124917#comment-17124917

Yun Tang commented on FLINK-10954:

I think multiple directories could benefit when using FsStateBackend if we have multi disks
and I believe this could make the disk usage more balance. On the other hand, if we decide
to choose one directory randomly as the base checkpoint directory of FsStateBackend, and there
exists many {{LocalRecoveryDirectoryProvider}}'s on that machine. I think this could achieve
similar effects compared with current implementation. I believe this could make the problem
much easier as you thought though this needs to refactor current interfaces of {{LocalRecoveryDirectoryProvider}}.


Besides, for RocksDB state-backend, we cannot just choose randomly as the base directory due
to recovery from previous local state also depends on the hard link mechanism.

> Hardlink from files of previous local stored state might cross devices
> ----------------------------------------------------------------------
>                 Key: FLINK-10954
>                 URL: https://issues.apache.org/jira/browse/FLINK-10954
>             Project: Flink
>          Issue Type: Bug
>          Components: Runtime / State Backends
>    Affects Versions: 1.6.2
>            Reporter: Yun Tang
>            Assignee: Yun Tang
>            Priority: Critical
>             Fix For: 1.12.0
> Currently, local recovery's base directories is initialized from '{{io.tmp.dirs}}' if
parameter '{{taskmanager.state.local.root-dirs}}' is not set. For Yarn environment, the tmp
dirs is replaced by its '{{LOCAL_DIRS}}', which might consist of directories from different
devices, such as /dump/1/nm-local-dir, /dump/2/nm-local-dir. The local directory for RocksDB
is initialized from IOManager's spillingDirectories, which might located in different device
from local recovery's folder. However, hard-link between different devices is not allowed,
it will throw exception below:
> {code:java}
> java.nio.file.FileSystemException: target -> souce: Invalid cross-device link
> {code}

This message was sent by Atlassian Jira

View raw message