[ 
https://issues.apache.org/jira/browse/CASSANDRA-17048?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17432909#comment-17432909
 ] 

Jacek Lewandowski commented on CASSANDRA-17048:
-----------------------------------------------

Basically I agree with you. Though, I don't understand the first point - the 
compatibility upgrades from 3.0 - what do you mean?

For mitigation options you mentioned, I think I like the option (1) the most. I 
don't know if we need to touch {{system}} keyspace during downgrading, the new 
table - {{sstable_activity_v2}} would just remain unused and the old stable 
{{sstable_activity}} would get repopulated. The tool would be super simple as 
it just needs to sort the sstables by generation ID and rename (perhaps 
snapshots as well). 

Option (2) feels tempting but without (1) it makes switching to the new scheme 
irreversible. Also it would introduce more complexity to the main code, while 
(1) keeps all the extra stuff needed for supporting downgrade in a separate 
tool.

I would not go for option (3) because of similar reasons the the mentioned 
above - extra complexity in the default path

For the patch complexity, git provides a bit misleading information that there 
is 100k+ changes - it is just because I added a new JMX dump. I'll rebase the 
PR and split it into reasonable chunks to make it more reviewer friendly.


> Replace sequential sstable generation identifier with ULID
> ----------------------------------------------------------
>
>                 Key: CASSANDRA-17048
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-17048
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Local/SSTable
>            Reporter: Jacek Lewandowski
>            Assignee: Jacek Lewandowski
>            Priority: Normal
>             Fix For: 4.1
>
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> Replace the current sequential sstable generation identifier with ULID based.
> ULID is better because we do not need to scan the existing files to pick the 
> starting number as well as we can generate globally unique identifiers. 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org

Reply via email to