Revert "Apply IndexBuildsOnJobFinishedId to unfinished builds only"

[?]

Mar 16, 2016, 4:04 PM

APURLT2Q3OKS6VF4HUWI3MGENXAKSSUDXI7CQOEGCRTYJRGUNT6QC

Dependencies

[2] K2WAI5SO Apply IndexBuildsOnJobFinishedId to unfinished builds only
[3] FANTYCR7 Materialize the number of finished builds
[4] KGGQD6CO Fix the Darcs schema upgrade
[5] 4S5JF5JP Use latest DBIx::Class::Schema::Loader
[6] 4CELXP7P Remove the longDescription field
[7] RJNMYHJF Add missing file
[8] PCKLFRT5 Support push notification of repository changes
[9] EJ7C77C3 Allow searching for store and drv paths. Only does a complete match to reduce load on database.
[10] GEADFVZ5 hydra-queue-runner: Improved scheduling
[11] LTQMSMYO Get rid of some obsolete indexes
[12] PHJF4FUT Add missing files
[13] H7HGWODE * Remove an empty file.
[14] RVHBV3B3 Notify the queue runner when a build is deleted
[15] PLOZBRTR Add command ‘hydra-queue-runner --status’ to show current status
[16] G2T4WAHI Store the inputs of each evaluation in the database
[17] QLOLZHRX Allow a per-jobset check interval
[18] CHJANKMM IndexJobsetEvalsOnJobsetId: Only index jobsets with new builds
[19] TM6C352U Ensure that stoptime and starttime are set for finished builds
[20] MOX7XJ2E Merge the BuildSchedulingInfo table into the Builds table
[21] Q5HZWFCY Add support for darcs repositories.
[22] P57OYJ4P Index builds on stop time
[23] TPNHTE5V Remove obsolete Builds columns and provide accurate "Running builds"
[24] DH3KNBAV Merge remote-tracking branch 'upstream/who-broke-builds' into upstream-master
[25] UNVMKJV5 Unify build and step status codes
[26] YU6CND7C Remove support for views
[27] D44B24QC Store the account type ("hydra" or "persona") explicitly in the database
[28] KDBWLKRP Index BuildSteps on propagatedFrom
[29] JD7AWXQG Remove now-unused SystemTypes table
[30] 2JGYLIEC Speed up the listing of jobset evals by caching some info in the DB
[31] E7FID72S Remove the BuildMachines and BuildMachinesSystemTypes tables
[32] UJOSXBZI Record which build a failed build step was propagated from
[33] 53IMJNBB Add isChannel column and meta attribute.
[34] ACBS7C6Q hydra-queue-runner: Detect changes to the scheduling shares
[35] BXHG3HYL When renaming a jobset, add a redirect from the old name
[36] 2NGV7NYN Drop unused "tag" column
[37] 4HPT4SDD Revert "Remove now-unused SystemTypes table"
[38] KN3VYE5P * Cleaned up the foreign key constraints.
[39] D5QIOJGP * Move everything up one directory.
[40] C5W3WZ3Y * Index the Builds table on drvPath to prevent a full table scan
[41] T5BIOVJE Add support for tracking custom metrics
[42] TX7Q4RAS Add page showing latest build steps
[43] KJQWSRCC Speed up hydra-update-gc-roots
[44] FV2M6MOT hydra: use autoconf/-make
[45] E7M2WP7A Remove unused Jobs columns
[46] 6QRHXIM3 * Speed up the jobset index page. Especially the query to get the
[47] FTPCV25M Store aggregate members in the database
[48] IWXLV4OB Add DB columns for when to notify responsible committers and which inputs should be checked
[49] URHBNFTI Add forgotten file
[50] WE5Q2NVI Allow build to be bumped to the front of the queue via the web interface
[51] 26PYNDXV Remove the errorMsg column from the Builds table
[52] FITVNQ2S Keep track of the time we spend copying to/from build machines
[53] SS4TZXNU Distinguish between permanent evaluation errors and transient input errors
[54] YTIDBFGU Drop unused "disabled" columns
[55] YF3HTIPF Use partial indexes
[56] RXVJFQ5A Evaluator cleanups
[57] KSBB33RE Add a dashboard
[58] W5OAZWPD Drop the errorMsg column in the Jobs table
[*] N22GPKYT * Put info about logs / build products in the DB.

Change contents

file deletion: upgrade-47.sql (----------)

[3.3004]→[2.281:319](∅→∅),[2.319]→[2.121:121](∅→∅)

drop index IndexBuildsOnJobFinishedId;
create index IndexBuildsOnJobFinishedId on builds(project, jobset, job, system, finished, id DESC) where finished = 0;
replacement in src/sql/hydra.sql at line 640

[3.79]→[2.0:119](∅→∅)

create index IndexBuildsOnJobFinishedId on builds(project, jobset, job, system, finished, id DESC) where finished = 0;

[3.79]

[3.0]

create index IndexBuildsOnJobFinishedId on builds(project, jobset, job, system, finished, id DESC);