[jira] [Commented] (ASTERIXDB-1289) Coordinate generated UUIDs within a cluster

["Yingyi Bu (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/ASTERIXDB-1289?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15137446#comment-15137446 ] 

Yingyi Bu commented on ASTERIXDB-1289:
--------------------------------------

This seems overlapping with what I'm working on, which is to add a function to generate unique ids for each tuple.  The new function is aim to fix ASTERIXDB-1168.

>> This could be improved by having the NCs obtain a unique pool from the CC to guarantee cluster-wide uniqueness.
The coordination seems not necessary. Once we migrate to the new function interface (IScalarEvaluator), a function is able to obtain its current "partition id" (which is cluster-wide unique for all partitions) from the IHyracksTaskContext.   


> Coordinate generated UUIDs within a cluster
> -------------------------------------------
>
>                 Key: ASTERIXDB-1289
>                 URL: https://issues.apache.org/jira/browse/ASTERIXDB-1289
>             Project: Apache AsterixDB
>          Issue Type: Bug
>            Reporter: Michael Blow
>   Original Estimate: 6h
>  Remaining Estimate: 6h
>
> UUIDs are generated by Asterix runtime for certain use cases; the uniqueness of the UUIDs relies on series of bytes obtained by SecureRandom on each NC.  This could be improved by having the NCs obtain a unique pool from the CC to guarantee cluster-wide uniqueness.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)