openstack-neutron launchpad start page timeout

[J.P.Klippel]

Everytime I open the openstack-neutron launchpad (https://launchpad.net/neutron) I get a timeout error.
The rest of the launchpad seems to work fine.

Here is an error id of a recent attempt: (Error ID: OOPS-ad67cca224bbbe6962ce96aa1a9c4869)

Is this the right forum to raise such a question?
Does this only affect me or does it affect anybody accessing the neutron-launchpad?
(The launchpads of other projects (e.g. openstack-nova) work fine).

[Colin Watson (cjwatson)]

Query plan on staging:

                                                                                                                                                         QUERY PLAN
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 Limit (cost=2205.94..2512.99 rows=5 width=984) (actual time=32.428..10994.752 rows=3 loops=1)
   Buffers: shared hit=339163 read=672
   CTE relevantspecification
     -> Index Scan using specification__product__name__key on specification specification_1 (cost=0.42..2180.05 rows=1286 width=4) (actual time=0.040..1.457 rows=1247 loops=1)
           Index Cond: (product = 31270)
           Buffers: shared hit=1243 read=10
   InitPlan 2 (returns $1)
     -> Aggregate (cost=4.12..4.13 rows=1 width=32) (actual time=0.048..0.049 rows=1 loops=1)
           Buffers: shared hit=3 read=1
           -> Index Scan using teamparticipation_person_idx on teamparticipation (cost=0.43..4.12 rows=2 width=4) (actual time=0.031..0.035 rows=12 loops=1)
                 Index Cond: (person = 4194263)
                 Buffers: shared hit=3 read=1
   InitPlan 3 (returns $3)
     -> Aggregate (cost=21.33..21.34 rows=1 width=32) (actual time=0.445..0.447 rows=1 loops=1)
           Buffers: shared hit=39 read=2
           -> Nested Loop (cost=0.85..21.33 rows=1 width=4) (actual time=0.433..0.435 rows=2 loops=1)
                 Buffers: shared hit=39 read=2
                 -> Index Scan using teamparticipation_person_idx on teamparticipation teamparticipation_1 (cost=0.43..4.12 rows=2 width=4) (actual time=0.007..0.011 rows=12 loops=1)
                       Index Cond: (person = 4194263)
                       Buffers: shared hit=4
                 -> Index Only Scan using accesspolicygrant__grantee__policy__key on accesspolicygrant (cost=0.42..8.57 rows=4 width=8) (actual time=0.034..0.035 rows=0 loops=12)
                       Index Cond: (grantee = teamparticipation_1.team)
                       Heap Fetches: 2
                       Buffers: shared hit=35 read=2
   -> Nested Loop Semi Join (cost=0.42..1634142.58 rows=26611 width=984) (actual time=32.427..10994.745 rows=3 loops=1)
         Join Filter: (specification.id = relevantspecification.id)
         Rows Removed by Join Filter: 52975988
         Buffers: shared hit=339163 read=672
         -> Index Scan Backward using specification__datecreated__id__idx on specification (cost=0.42..94151.15 rows=53222 width=984) (actual time=0.045..423.806 rows=42484 loops=1)
               Filter: ((implementation_status <> 90) AND (definition_status <> ALL ('{70,60}'::integer[])) AND ((implementation_status <> 95) OR (definition_status <> 10)) AND ((information_type = ANY ('{1,2}'::integer[])) OR COALESCE((access_grants && $1), false) OR COALESCE((ARRAY[access_policy] && $3), false)))
               Rows Removed by Filter: 298230
               Buffers: shared hit=337920...

[Jürgen Gmach (jugmac00)]

My colleague has identified the reason and is preparing a fix.

[Colin Watson (cjwatson)]

We've optimized this page and it now performs acceptably.

[J.P.Klippel (fbccfba67e05)]

Can confirm. Thanks for your effort!