[
https://issues.apache.org/jira/browse/KYLIN-6017?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17928283#comment-17928283
]
Guoliang Sun commented on KYLIN-6017:
-------------------------------------
h3. Root Cause
The metadata recovery process is divided into several steps, wrapped in a large
Transaction:
# Traverse `resourceUrl`, write to the database, and update the cache.
# After the Transaction commit, there is a follow-up `endTransaction`
operation:
## Record the audit log of this transaction and write it to the `auditLog`
table.
## Replay the audit log for upsert or delete operations.
Due to the default timeout mechanism of
`kylin.metadata.audit-log.catchup-max-timeout=60s`, replaying the audit log
will time out when the metadata being restored is too large.
> Metadata recovery times out when the metadata size is too large
> ---------------------------------------------------------------
>
> Key: KYLIN-6017
> URL: https://issues.apache.org/jira/browse/KYLIN-6017
> Project: Kylin
> Issue Type: Bug
> Affects Versions: 5.0.0
> Reporter: Guoliang Sun
> Priority: Major
> Attachments: image-2025-02-19-14-45-02-996.png
>
>
> !image-2025-02-19-14-45-02-996.png|width=640,height=189!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
