Ensembl/ensembl-hive

View on GitHub
sql/procedures.sqlite

Summary

Maintainability
Test Coverage
/*

DESCRIPTION

    Some stored functions, views and procedures used in eHive


LICENSE

    Copyright [1999-2015] Wellcome Trust Sanger Institute and the EMBL-European Bioinformatics Institute
    Copyright [2016-2021] EMBL-European Bioinformatics Institute

    Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License.
    You may obtain a copy of the License at

         http://www.apache.org/licenses/LICENSE-2.0

    Unless required by applicable law or agreed to in writing, software distributed under the License
    is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    See the License for the specific language governing permissions and limitations under the License.

CONTACT

    Please subscribe to the Hive mailing list:  http://listserver.ebi.ac.uk/mailman/listinfo/ehive-users  to discuss Hive-related questions or to be notified of our updates

*/


-- show hive progress for analyses (turned into a view to give extra flexibility) ----------------
--
-- Thanks to Greg Jordan for the idea and the original version
--
-- Usage:
--       select * from progress;                                         # the whole table (may take ages to generate, depending on the size of your pipeline)
--       select * from progress where logic_name like 'family_blast%';   # only show family_blast-related analyses
--       select * from progress where retry_count>1;                     # only show jobs that have been tried more than once

CREATE VIEW IF NOT EXISTS progress AS
    SELECT a.logic_name || '(' || a.analysis_id || ')' analysis_name_and_id,
        MIN(rc.name) resource_class,
        j.status,
        j.retry_count,
        CASE WHEN j.status IS NULL THEN 0 ELSE count(*) END cnt,
        MIN(job_id) example_job_id
    FROM        analysis_base a
    LEFT JOIN   job j USING (analysis_id)
    LEFT JOIN   resource_class rc ON (a.resource_class_id=rc.resource_class_id)
    GROUP BY a.analysis_id, j.status, j.retry_count
    ORDER BY a.analysis_id, j.status;


-- a convenient view that also incorporates (otherwise redundant) analysis_id and logic_name ------------
--
-- Usage:
--       select * from msg;
--       select * from msg where analysis_id=18;
--       select * from msg where logic_name like 'family_blast%';

CREATE VIEW IF NOT EXISTS msg AS
    SELECT a.analysis_id, a.logic_name, m.*
    FROM log_message m
    LEFT JOIN role USING (role_id)
    LEFT JOIN analysis_base a USING (analysis_id);


-- show statistics of Workers' real resource usage by analysis -------------------------------------------
--
-- Usage:
--       select * from resource_usage_stats;
--       select * from resource_usage_stats where logic_name like 'family_blast%';

CREATE VIEW IF NOT EXISTS resource_usage_stats AS
    SELECT a.logic_name || '(' || a.analysis_id || ')' analysis,
           w.meadow_type,
           rc.name || '(' || rc.resource_class_id || ')' resource_class,
           u.exit_status,
           count(*) workers,
           min(mem_megs) AS min_mem_megs, round(avg(mem_megs),2) AS avg_mem_megs, max(mem_megs) AS max_mem_megs,
           min(swap_megs) AS min_swap_megs, round(avg(swap_megs),2) AS avg_swap_megs, max(swap_megs) AS max_swap_megs
    FROM resource_class rc
    JOIN analysis_base a USING(resource_class_id)
    LEFT JOIN role r USING(analysis_id)
    LEFT JOIN worker w USING(worker_id)
    LEFT JOIN worker_resource_usage u USING (worker_id)
    GROUP BY a.analysis_id, w.meadow_type, rc.resource_class_id, u.exit_status
    ORDER BY a.analysis_id, w.meadow_type, rc.resource_class_id, u.exit_status;


-- show the roles that are currently live (grouped by meadow_users, resource_classes and analyses) -------
--
-- Usage:
--       select * from live_roles;
--       select * from live_roles where resource_class_id=12;

CREATE VIEW IF NOT EXISTS live_roles AS
    SELECT w.meadow_user, w.meadow_type, w.resource_class_id, rc.name resource_class_name, r.analysis_id, a.logic_name, count(*)
    FROM worker w
    JOIN role r USING(worker_id)
    LEFT JOIN resource_class rc ON w.resource_class_id=rc.resource_class_id
    LEFT JOIN analysis_base a ON a.analysis_id = r.analysis_id
    WHERE r.when_finished IS NULL
    GROUP BY w.meadow_user, w.meadow_type, w.resource_class_id, rc.name, r.analysis_id, a.logic_name;