Use postgresql recursive feature for getting scheduled product

[perlpunk]

For longer chains of clones it's more efficient. It also avoids a warning:

Deep recursion on subroutine "OpenQA::Schema::Result::Jobs::related_scheduled_product_id"

Issue: https://progress.opensuse.org/issues/152569

[kalikiana]

For longer chains of clones it's more efiicient. It also avoids a warning:

Typo: efficient

The SQL looks correct, but I'm not sure how to tell if it's more or less efficient.

[perlpunk]

Regarding efficiency:

openqa=> explain analyze with recursive orig_id as (
            select (13062217)::bigint as orig_id, (null)::bigint as prod
            union all
select id as orig_id, scheduled_product_id as prod from jobs join orig_id on orig_id.orig_id = jobs.clone_id)
select orig_id, prod from orig_id where prod is not null order by orig_id desc;
                                                                   QUERY PLAN                                                                   
------------------------------------------------------------------------------------------------------------------------------------------------
 Sort  (cost=854.62..854.87 rows=100 width=16) (actual time=10.769..10.770 rows=1 loops=1)
   Sort Key: orig_id.orig_id DESC
   Sort Method: quicksort  Memory: 25kB
   CTE orig_id
     ->  Recursive Union  (cost=0.00..849.28 rows=101 width=16) (actual time=0.002..10.420 rows=1099 loops=1)
           ->  Result  (cost=0.00..0.01 rows=1 width=16) (actual time=0.001..0.001 rows=1 loops=1)
           ->  Nested Loop  (cost=0.42..84.72 rows=10 width=16) (actual time=0.009..0.009 rows=1 loops=1099)
                 ->  WorkTable Scan on orig_id orig_id_1  (cost=0.00..0.20 rows=10 width=8) (actual time=0.000..0.000 rows=1 loops=1099)
                 ->  Index Scan using jobs_idx_clone_id on jobs  (cost=0.42..8.44 rows=1 width=24) (actual time=0.008..0.008 rows=1 loops=1099)
                       Index Cond: (clone_id = orig_id_1.orig_id)
   ->  CTE Scan on orig_id  (cost=0.00..2.02 rows=100 width=16) (actual time=10.757..10.764 rows=1 loops=1)
         Filter: (prod IS NOT NULL)
         Rows Removed by Filter: 1098
 Planning Time: 0.187 ms
 Execution Time: 10.811 ms
(15 rows)

And that example is a job in osd which has about 1100 ancestors.

[perlpunk]

And we're using that kind of SQL already in ancestors and descendants

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (e68161c) 98.36% compared to head (e56b56e) 98.37%.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #5398   +/-   ##
=======================================
  Coverage   98.36%   98.37%           
=======================================
  Files         389      389           
  Lines       37534    37548   +14     
=======================================
+ Hits        36922    36936   +14     
  Misses        612      612           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.