Xiao-zhen-Liu opened a new pull request, #8734:
URL: https://github.com/apache/texera/pull/8734

   ### What changes were proposed in this PR?
   
   This PR is part of cache reuse (#5881), where a run reads the saved results 
of an earlier run instead of computing them again. When 
`FriesReconfigurationAlgorithm` works out which operators a reconfiguration 
must pause, it now takes an operator's parents and children from the region's 
links instead of the operator's own link lists.
   
   - On main every link between two operators of a region is one of the 
region's links, so every reconfiguration keeps the same scope.
   - With cache reuse, an operator can read a saved port whose own operator 
also runs. For example, Split runs for U2 through its port 1 while U1 reads 
Split's port 0 from its saved result. The plan that gets scheduled leaves the 
link from Split to U1 out (#8733), but Split and U1 still list it on their 
ports. Following the operators' own lists, reconfiguring Split and U2 would 
also pause U1, which gets no data from Split in that run.
   
   ### Any related issues, documentation, discussions?
   
   Part of #5881 (cache reuse). Advances #5884. Design discussion: #5880. 
Related: #8733, which schedules a run with cache reuse and leaves such links 
out of the scheduled plan.
   
   ### How was this PR tested?
   
   `FriesReconfigurationAlgorithmSpec` gains two cases in which a link that 
only the operators of a region list would otherwise add an operator to the 
reconfiguration's scope. All specs of the scheduling, coordinator and 
`engine.common` packages and `scalafmtCheckAll` pass.
   
   ### Was this PR authored or co-authored using generative AI tooling?
   
   Generated-by: Claude (Claude Code)
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to