[jira] [Comment Edited] (IGNITE-16501) TX Recovery protocol in Cockroach in case the tx coordinator failure

["Sergey Uttsel (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IGNITE-16501?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17496057#comment-17496057 ] 

Sergey Uttsel edited comment on IGNITE-16501 at 3/23/22, 7:54 AM:
------------------------------------------------------------------

https://issues.apache.org/jira/browse/IGNITE-16723 was created to investigate other tx recovery cases.


was (Author: sergey uttsel):
After investigation the process of transaction recovery in the STAGED abandoned status, the following questions remained:
Where are intents stored before they are replicated? If in the main storage, then how do they get there if they have not yet been replicated?
How is it checked that the intent is replicated?

> TX Recovery protocol in Cockroach in case the tx coordinator failure
> --------------------------------------------------------------------
>
>                 Key: IGNITE-16501
>                 URL: https://issues.apache.org/jira/browse/IGNITE-16501
>             Project: Ignite
>          Issue Type: Task
>            Reporter: Vyacheslav Koptilin
>            Assignee: Sergey Uttsel
>            Priority: Major
>              Labels: ignite-3
>         Attachments: Cockroach transaction status recovery.pdf
>
>
> The flowchart [^Cockroach transaction status recovery.pdf] describe transaction recovery if txn coordinator was failed.
> Need to investigate other failures: a txn record leaseholder fail, a write intent leaseholder fail and etc.



--
This message was sent by Atlassian Jira
(v8.20.1#820001)