RunCommandLogs: init table

[?]

Nov 17, 2021, 5:34 PM

2GNJU4VFSYWUXS3JG3XZZFM3QSN3ILOIJVJFA75TKBIFS2WFNXBQC

Dependencies

[2] 6JYDU2DM TaskRetries: init table
[3] 3ZCEPLNO
[4] BD3GRK4B * Get rid of "positive failures" and separate log phases.
[5] BHZXGT2H * Channels: provide an index page that lists all the packages in the
[6] H7CNGK4O * Log evaluation errors etc. in the DB.
[7] L2E6EVE2 * Merged the Build and Job tables.
[8] TWVSALRL * Allow the maximum number of concurrent builds per platform to be
[9] 3E6IP3R3 * Add the name of the jobset to ReleaseSetJobs, otherwise we can't
[10] YTIDBFGU Drop unused "disabled" columns
[11] V4RNHJNR * Add a link to each project's homepage. Suggested by karltk.
[12] TLZ2SPBR
[13] ITJWNQEZ Update Hydra schema, otherwise hydra-notify will not work.
[14] ECBA3GQO * Make the schema class names match the case of the SQL table names.
[15] G2T4WAHI Store the inputs of each evaluation in the database
[16] B72GLND4
[17] ICRGXEPI Builds.jobset_id: make not-null
[18] SJLEZFC4 check getHydraPath in stead of Envvar HYDRA_DBI directly
[19] QUMWPGCU Add flake configuration to the web interface
[20] TPNHTE5V Remove obsolete Builds columns and provide accurate "Running builds"
[21] SHBLLAVH * More global substitution.
[22] D3DIBMOK * For products that are directories (like manuals), allow a default
[23] Y6AHH4TH Remove the logfile and logSize columns from the database
[24] MB5INTIQ Normalize nixexpr{input,path} from builds to jobsetevals.
[25] AHTEIK7G * Added a maintainers field to the Builds table.
[26] 4S5JF5JP Use latest DBIx::Class::Schema::Loader
[27] S5PV6IIM * Represent jobs explicitly in the DB.
[28] WE5Q2NVI Allow build to be bumped to the front of the queue via the web interface
[29] JTHWA6AM Rename aggregate members to constituents
[30] PMNWRTGJ Add multiple output support
[31] YTZOC7C5 * Editing of jobset inputs.
[32] IK53RV4V
[33] IGR322YP sql: Generate models from postgresql
[34] 6BLUKEQ2 * Caching of "path" inputs, and fake a revision number for those.
[35] GNIEG2GC * Disambiguate jobs by jobset name. I.e. jobs with the same name in
[36] RU7AQO7U * Role-based access control. Only admins can create projects. Only
[37] TQKGQ5R3
[38] FTPCV25M Store aggregate members in the database
[39] WQ3QMV3D DBIx::Class: migrate to use_namespaces
[40] YAPITGB3 * Boolean inputs.
[41] LZVO64YG Merge in the first bits of the API work
[42] T5BIOVJE Add support for tracking custom metrics
[43] EYNG4EL4 * Regenerate the bindings from a clean sqlite database.
[44] 4CELXP7P Remove the longDescription field
[45] R5D7DZPE
[46] LCKWLQW3 * In Sqlite "release" is now a keyword, so use "release_" instead.
[47] S66BOMVU * Added authentication.
[48] ZI535LI6 * hydra: 'new' UI for project/jobset/job/build
[49] ZVTSOVHN * Support Subversion checkouts.
[50] SJPEC375 Builds: add a nullable jobset_id foreign key to Jobsets.
[51] CLJQCY2X * Store info about all the build actions and allow them to be
[52] VJHIHMEH * Store the meta.longDescription and meta.license attributes in the
[53] KOTB7BKV
[54] 26PYNDXV Remove the errorMsg column from the Builds table
[55] P3M6UFMP Merge remote-tracking branch 'origin/master' into flake
[56] FHF6IZJQ * Basic release management: releases are now dynamically computed as
[57] 53IMJNBB Add isChannel column and meta attribute.
[58] N74SRHS6 Remove the "releases" feature
[59] NZXX6PLD Remove the Jobs table
[60] CMU3YKOU * Store the release name.
[61] 37R34XJO * Negative caching: don't perform a build if a dependency already
[*] J5UVLXOK * Start of a basic Catalyst web interface.
[*] RFE6T5LG * Store jobset evaluations in the database explicitly. This includes
[*] N22GPKYT * Put info about logs / build products in the DB.
[*] RYTQLATY Keep track of failed paths in the Hydra database
[*] OZ2XXTT4 update-dbix.pl: correct indentation

Change contents

edit in src/lib/Hydra/Schema/Result/Builds.pm at line 501

[3.3422]

[64.1050]

=head2 runcommandlogs

Type: has_many

Related object: L<Hydra::Schema::Result::RunCommandLogs>
edit in src/lib/Hydra/Schema/Result/Builds.pm at line 508

[64.1051]

[3.3976]

=cut

__PACKAGE__->has_many(
"runcommandlogs",
"Hydra::Schema::Result::RunCommandLogs",
{ "foreign.build_id" => "self.id" },
undef,
);
replacement in src/lib/Hydra/Schema/Result/Builds.pm at line 546

[3.3378]→[3.4744:4886](∅→∅)

# Created by DBIx::Class::Schema::Loader v0.07049 @ 2021-08-26 12:02:36
# DO NOT MODIFY THIS OR ANYTHING ABOVE! md5sum:WHdSVHhQykmUz0tR/TExVg

[3.3378]

[3.4481]

# Created by DBIx::Class::Schema::Loader v0.07049 @ 2021-11-17 12:42:34
# DO NOT MODIFY THIS OR ANYTHING ABOVE! md5sum:ylttv/NTMDcSZumBXRCOCw
file addition: RunCommandLogs.pm (----------)

[3.1]

use utf8;
package Hydra::Schema::Result::RunCommandLogs;

# Created by DBIx::Class::Schema::Loader
# DO NOT MODIFY THE FIRST PART OF THIS FILE

=head1 NAME

Hydra::Schema::Result::RunCommandLogs

=cut

use strict;
use warnings;

use base 'DBIx::Class::Core';

=head1 COMPONENTS LOADED

=over 4

=item * L<Hydra::Component::ToJSON>

=back

=cut

__PACKAGE__->load_components("+Hydra::Component::ToJSON");

=head1 TABLE: C<runcommandlogs>

=cut

__PACKAGE__->table("runcommandlogs");

=head1 ACCESSORS

=head2 id

data_type: 'integer'
is_auto_increment: 1
is_nullable: 0
sequence: 'runcommandlogs_id_seq'

=head2 job_matcher

data_type: 'text'
is_nullable: 0

=head2 build_id

data_type: 'integer'
is_foreign_key: 1
is_nullable: 0

=head2 command

data_type: 'text'
is_nullable: 0

=head2 start_time

data_type: 'integer'
is_nullable: 1

=head2 end_time

data_type: 'integer'
is_nullable: 1

=head2 error_number

data_type: 'integer'
is_nullable: 1

=head2 exit_code

data_type: 'integer'
is_nullable: 1

=head2 signal

data_type: 'integer'
is_nullable: 1

=head2 core_dumped

data_type: 'boolean'
is_nullable: 1

=cut

__PACKAGE__->add_columns(
"id",
{
data_type => "integer",
is_auto_increment => 1,
is_nullable => 0,
sequence => "runcommandlogs_id_seq",
},
"job_matcher",
{ data_type => "text", is_nullable => 0 },
"build_id",
{ data_type => "integer", is_foreign_key => 1, is_nullable => 0 },
"command",
{ data_type => "text", is_nullable => 0 },
"start_time",
{ data_type => "integer", is_nullable => 1 },
"end_time",
{ data_type => "integer", is_nullable => 1 },
"error_number",
{ data_type => "integer", is_nullable => 1 },
"exit_code",
{ data_type => "integer", is_nullable => 1 },
"signal",
{ data_type => "integer", is_nullable => 1 },
"core_dumped",
{ data_type => "boolean", is_nullable => 1 },
);

=head1 PRIMARY KEY

=over 4

=item * L</id>

=back

=cut

__PACKAGE__->set_primary_key("id");

=head1 RELATIONS

=head2 build

Type: belongs_to

Related object: L<Hydra::Schema::Result::Builds>

=cut

__PACKAGE__->belongs_to(
"build",
"Hydra::Schema::Result::Builds",
{ id => "build_id" },
{ is_deferrable => 0, on_delete => "CASCADE", on_update => "NO ACTION" },
);

# Created by DBIx::Class::Schema::Loader v0.07049 @ 2021-11-18 12:35:52
# DO NOT MODIFY THIS OR ANYTHING ABOVE! md5sum:N0G71diB8DNDgkYgaSQrFA

# You can replace this text with custom code or comments, and it will be preserved on regeneration
1;
edit in src/sql/hydra.sql at line 564

[2.2328]

[66.4073]
edit in src/sql/hydra.sql at line 566

[66.4074]

[66.4074]

-- Records of RunCommand executions
--
-- The intended flow is:
--
-- 1. Create a RunCommandLogs entry when the task is "queued" to run
-- 2. Update the start_time when it begins
-- 3. Update the end_time and exit_code when it completes
create table RunCommandLogs (
id serial primary key not null,
job_matcher text not null,
build_id integer not null,
-- TODO: evaluation_id integer not null,
-- can we do this in a principled way? a build can be part of many evaluations
-- but a "bug" of RunCommand, imho, is that it should probably run per evaluation?
command text not null,
start_time integer,
end_time integer,
error_number integer,
exit_code integer,
signal integer,
core_dumped boolean,

foreign key (build_id) references Builds(id) on delete cascade,
-- foreign key (evaluation_id) references Builds(id) on delete cascade,

constraint RunCommandLogs_not_started_no_exit_time_no_code check (
-- If start time is null, then end_time, exit_code, signal, and core_dumped should be null.
-- A logical implication operator would be nice :).
(start_time is not null) or (
end_time is null
and error_number is null
and exit_code is null
and signal is null
and core_dumped is null
)
),
constraint RunCommandLogs_end_time_has_start_time check (
-- If end time is not null, then end_time, exit_code, and core_dumped should not be null
(end_time is null) or (start_time is not null)
)

-- Note: if exit_code is not null then signal and core_dumped must be null.
-- Similarly, if signal is not null then exit_code must be null and
-- core_dumped must not be null. However, these semantics are tricky
-- to encode as constraints and probably provide limited actual value.
);
edit in src/sql/update-dbix.pl at line 41

[67.1194]

[67.1194]

"runcommandlogs" => "RunCommandLogs",