tables.sqlite 29.6 KB
Newer Older
1 2
/*
    This is SQLite version of EnsEMBL Hive database schema file.
Leo Gordon's avatar
Leo Gordon committed
3

4 5
    It has been annotated with @-tags.
    The following command is used to create HTML documentation:
Leo Gordon's avatar
Leo Gordon committed
6 7 8 9 10
        perl $ENSEMBL_CVS_ROOT_DIR/ensembl-production/scripts/sql2html.pl \
            -i $ENSEMBL_CVS_ROOT_DIR/ensembl-hive/sql/tables.sqlite -d Hive -intro $ENSEMBL_CVS_ROOT_DIR/ensembl-hive/docs/hive_schema.inc \
            -sort_headers 0 -sort_tables 0 -o $ENSEMBL_CVS_ROOT_DIR/ensembl-hive/docs/hive_schema.html


11 12 13
    Adding the following line into the header of the previous output will make it look prettier:
            <link rel="stylesheet" type="text/css" media="all" href="ehive_doc.css" />

14

15 16 17 18


LICENSE

19
    Copyright [1999-2015] Wellcome Trust Sanger Institute and the EMBL-European Bioinformatics Institute
nwillhoft's avatar
nwillhoft committed
20
    Copyright [2016-2021] EMBL-European Bioinformatics Institute
21 22 23 24 25 26 27 28 29 30

    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.

31 32 33 34
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

35 36 37 38 39 40 41 42
*/


/**
@header Pipeline structure
@colour #C70C09
*/

Leo Gordon's avatar
Leo Gordon committed
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81
/**
@table  hive_meta

@colour #000000

@desc This table keeps several important hive-specific pipeline-wide key-value pairs
        such as hive_sql_schema_version, hive_use_triggers and hive_pipeline_name.

@column meta_key        the KEY of KEY-VALUE pairs (primary key)
@column meta_value      the VALUE of KEY-VALUE pairs
*/

CREATE TABLE hive_meta (
    meta_key                VARCHAR(255) NOT NULL PRIMARY KEY,
    meta_value              TEXT

);


/**
@table  pipeline_wide_parameters

@colour #000000

@desc This table contains a simple hash between pipeline_wide_parameter names and their values.
      The same data used to live in 'meta' table until both the schema and the API were finally separated from Ensembl Core.

@column param_name      the KEY of KEY-VALUE pairs (primary key)
@column param_value     the VALUE of KEY-VALUE pairs
*/

CREATE TABLE pipeline_wide_parameters (
    param_name              VARCHAR(255) NOT NULL PRIMARY KEY,
    param_value             TEXT

);
CREATE        INDEX pipeline_wide_parameters_value_idx     ON pipeline_wide_parameters (param_value);


82 83 84 85 86 87 88 89 90 91 92
/**
@table  analysis_base

@colour #C70C09

@desc   Each Analysis is a node of the pipeline diagram.
        It acts both as a "class" to which Jobs belong (and inherit from it certain properties)
        and as a "container" for them (Jobs of an Analysis can be blocking all Jobs of another Analysis).

@column analysis_id             a unique ID that is also a foreign key to most of the other tables
@column logic_name              the name of the Analysis object
Matthieu Muffato's avatar
Matthieu Muffato committed
93
@column module                  the name of the module / package that runs this Analysis
94
@column language                the language of the module, if not Perl
95 96 97 98 99 100 101 102 103
@column parameters              a stingified hash of parameters common to all jobs of the Analysis
@column resource_class_id       link to the resource_class table
@column failed_job_tolerance    % of tolerated failed Jobs
@column max_retry_count         how many times a job of this Analysis will be retried (unless there is no point)
@column can_be_empty            if TRUE, this Analysis will not be blocking if/while it doesn't have any jobs
@column priority                an Analysis with higher priority will be more likely chosen on Worker's specialization
@column meadow_type             if defined, forces this Analysis to be run only on the given Meadow
@column analysis_capacity       if defined, limits the number of Workers of this particular Analysis that are allowed to run in parallel
*/
104

105
CREATE TABLE analysis_base (
Leo Gordon's avatar
Leo Gordon committed
106
    analysis_id             INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
107
    logic_name              VARCHAR(255) NOT NULL,
108
    module                  VARCHAR(255) NOT NULL,
109
    language                VARCHAR(255),
110
    parameters              TEXT,
Leo Gordon's avatar
Leo Gordon committed
111 112 113 114 115
    resource_class_id       INTEGER     NOT NULL,
    failed_job_tolerance    INTEGER     NOT NULL DEFAULT 0,
    max_retry_count         INTEGER     NOT NULL DEFAULT 3,
    can_be_empty            SMALLINT    NOT NULL DEFAULT 0,
    priority                SMALLINT    NOT NULL DEFAULT 0,
116
    meadow_type             VARCHAR(255)         DEFAULT NULL,
Leo Gordon's avatar
Leo Gordon committed
117
    analysis_capacity       INTEGER              DEFAULT NULL
118
);
Leo Gordon's avatar
Leo Gordon committed
119
CREATE UNIQUE INDEX analysis_base_logic_name_idx ON analysis_base (logic_name);
120 121


122 123
/**
@table  analysis_stats
124

125
@colour #C70C09
126

127 128 129 130
@desc   Parallel table to analysis_base which provides high level statistics on the
        state of an analysis and it's jobs.  Used to provide a fast overview, and to
        provide final approval of 'DONE' which is used by the blocking rules to determine
        when to unblock other analyses.  Also provides
131

132 133 134 135
@column analysis_id             foreign-keyed to the corresponding analysis_base entry
@column batch_size              how many jobs are claimed in one claiming operation before Worker starts executing them
@column hive_capacity           a reciprocal limiter on the number of Workers running at the same time (dependent on Workers of other Analyses)
@column status                  cached state of the Analysis
136

137 138 139 140 141 142 143
@column total_job_count         total number of Jobs of this Analysis
@column semaphored_job_count    number of Jobs of this Analysis that are in SEMAPHORED state
@column ready_job_count         number of Jobs of this Analysis that are in READY state
@column done_job_count          number of Jobs of this Analysis that are in DONE state
@column failed_job_count        number of Jobs of this Analysis that are in FAILED state

@column num_running_workers     number of running Workers of this Analysis
144

145 146 147
@column behaviour               whether hive_capacity is set or is dynamically calculated based on timers
@column input_capacity          used to compute hive_capacity in DYNAMIC mode
@column output_capacity         used to compute hive_capacity in DYNAMIC mode
148

149 150 151 152 153
@column avg_msec_per_job        weighted average used to compute DYNAMIC hive_capacity
@column avg_input_msec_per_job  weighted average used to compute DYNAMIC hive_capacity
@column avg_run_msec_per_job    weighted average used to compute DYNAMIC hive_capacity
@column avg_output_msec_per_job weighted average used to compute DYNAMIC hive_capacity

154
@column when_updated            when this entry was last updated
155 156 157 158
@column sync_lock               a binary lock flag to prevent simultaneous updates
*/

CREATE TABLE analysis_stats (
Leo Gordon's avatar
Leo Gordon committed
159 160 161 162 163 164 165 166 167 168
    analysis_id             INTEGER     NOT NULL,
    batch_size              INTEGER     NOT NULL DEFAULT 1,
    hive_capacity           INTEGER              DEFAULT NULL,
    status                  TEXT        NOT NULL DEFAULT 'EMPTY', /* enum('BLOCKED', 'LOADING', 'SYNCHING', 'EMPTY', 'READY', 'WORKING', 'ALL_CLAIMED', 'DONE', 'FAILED') DEFAULT 'EMPTY' NOT NULL, */

    total_job_count         INTEGER     NOT NULL DEFAULT 0,
    semaphored_job_count    INTEGER     NOT NULL DEFAULT 0,
    ready_job_count         INTEGER     NOT NULL DEFAULT 0,
    done_job_count          INTEGER     NOT NULL DEFAULT 0,
    failed_job_count        INTEGER     NOT NULL DEFAULT 0,
169

Leo Gordon's avatar
Leo Gordon committed
170 171 172 173 174 175 176 177 178 179 180
    num_running_workers     INTEGER     NOT NULL DEFAULT 0,

    behaviour               TEXT        NOT NULL DEFAULT 'STATIC', /* enum('STATIC', 'DYNAMIC') DEFAULT 'STATIC' NOT NULL, */
    input_capacity          INTEGER     NOT NULL DEFAULT 4,
    output_capacity         INTEGER     NOT NULL DEFAULT 4,

    avg_msec_per_job        INTEGER              DEFAULT NULL,
    avg_input_msec_per_job  INTEGER              DEFAULT NULL,
    avg_run_msec_per_job    INTEGER              DEFAULT NULL,
    avg_output_msec_per_job INTEGER              DEFAULT NULL,

181
    when_updated            TIMESTAMP            DEFAULT NULL,
Leo Gordon's avatar
Leo Gordon committed
182
    sync_lock               SMALLINT    NOT NULL DEFAULT 0,
183 184 185 186 187 188 189 190 191 192

    PRIMARY KEY (analysis_id)
);


/**
@table  dataflow_rule

@colour #C70C09

193 194 195 196
@desc Each entry of this table defines a starting point for dataflow (via from_analysis_id and branch_code)
      to which point a group of dataflow_target entries can be linked. This grouping is used in two ways:
      (1) dataflow_target entries that link into the same dataflow_rule share the same from_analysis, branch_code and funnel_dataflow_rule
      (2) to define the conditions for DEFAULT or ELSE case (via excluding all conditions explicitly listed in the group)
Leo Gordon's avatar
Leo Gordon committed
197 198 199 200 201

@column dataflow_rule_id        internal ID
@column from_analysis_id        foreign key to analysis table analysis_id
@column branch_code             branch_code of the fan
@column funnel_dataflow_rule_id dataflow_rule_id of the semaphored funnel (is NULL by default, which means dataflow is not semaphored)
202
*/
203

204
CREATE TABLE dataflow_rule (
Leo Gordon's avatar
Leo Gordon committed
205 206 207
    dataflow_rule_id        INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
    from_analysis_id        INTEGER     NOT NULL,
    branch_code             INTEGER     NOT NULL DEFAULT 1,
208 209 210 211 212 213 214 215 216 217 218
    funnel_dataflow_rule_id INTEGER              DEFAULT NULL
);


/**
@table  dataflow_target

@colour #C70C09

@desc This table links specific conditions with the target object (Analysis/Table/Accu) and optional input_id_template.

219
@column dataflow_target_id      internal ID
220 221 222
@column source_dataflow_rule_id foreign key to the dataflow_rule object that defines grouping (see description of dataflow_rule table)
@column on_condition            param-substitutable string evaluated at the moment of dataflow event that defines whether or not this case produces any dataflow; NULL means DEFAULT or ELSE
@column input_id_template       a template for generating a new input_id (not necessarily a hashref) in this dataflow; if undefined is kept original
223
@column extend_param_stack      the boolean value defines whether the newly created jobs will inherit both the parameters and the accu of the prev_job
224 225 226 227
@column to_analysis_url         the URL of the dataflow target object (Analysis/Table/Accu)
*/

CREATE TABLE dataflow_target (
228
    dataflow_target_id      INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
229 230 231
    source_dataflow_rule_id INTEGER     NOT NULL,
    on_condition            VARCHAR(255)          DEFAULT NULL,
    input_id_template       TEXT                  DEFAULT NULL,
232
    extend_param_stack      SMALLINT     NOT NULL DEFAULT 0,
233
    to_analysis_url         VARCHAR(255) NOT NULL DEFAULT ''       -- to be renamed 'target_url'
234
);
235
CREATE UNIQUE INDEX all_of_dataflow_target_idx ON dataflow_target (source_dataflow_rule_id, on_condition, to_analysis_url, input_id_template);
236 237


238 239 240 241 242 243 244 245 246 247 248 249
/**
@table  analysis_ctrl_rule

@colour #C70C09

@desc   These rules define a higher level of control.
        These rules are used to turn whole anlysis nodes on/off (READY/BLOCKED).
        If any of the condition_analyses are not 'DONE' the ctrled_analysis is set to BLOCKED.
        When all conditions become 'DONE' then ctrled_analysis is set to READY
        The workers switch the analysis.status to 'WORKING' and 'DONE'.
        But any moment if a condition goes false, the analysis is reset to BLOCKED.

250
@column analysis_ctrl_rule_id  internal ID
251 252 253
@column condition_analysis_url foreign key to net distributed analysis reference
@column ctrled_analysis_id     foreign key to analysis table analysis_id
*/
254 255

CREATE TABLE analysis_ctrl_rule (
256
    analysis_ctrl_rule_id   INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
Leo Gordon's avatar
Leo Gordon committed
257 258
    condition_analysis_url  VARCHAR(255) NOT NULL DEFAULT '',
    ctrled_analysis_id      INTEGER     NOT NULL
259
);
Leo Gordon's avatar
Leo Gordon committed
260
CREATE UNIQUE INDEX condition_ctrled_idx ON analysis_ctrl_rule (condition_analysis_url, ctrled_analysis_id);
261 262


263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279
/**
@header Resources
@colour #FF7504
*/

/**
@table  resource_class

@colour #FF7504

@desc   Maps between resource_class numeric IDs and unique names.

@column resource_class_id   unique ID of the ResourceClass
@column name                unique name of the ResourceClass
*/

CREATE TABLE resource_class (
Leo Gordon's avatar
Leo Gordon committed
280
    resource_class_id       INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
281
    name                    VARCHAR(255) NOT NULL
282 283

);
Leo Gordon's avatar
Leo Gordon committed
284
CREATE UNIQUE INDEX resource_class_name_idx ON resource_class (name);
285 286 287 288 289 290 291 292 293 294 295


/**
@table  resource_description

@colour #FF7504

@desc   Maps (ResourceClass, MeadowType) pair to Meadow-specific resource lines.

@column resource_class_id   foreign-keyed to the ResourceClass entry
@column meadow_type         if the Worker is about to be executed on the given Meadow...
296 297
@column submission_cmd_args ... these are the resource arguments (queue, memory,...) to give to the submission command
@column worker_cmd_args     ... and these are the arguments that are given to the worker command being submitted
298 299 300
*/

CREATE TABLE resource_description (
Leo Gordon's avatar
Leo Gordon committed
301
    resource_class_id       INTEGER     NOT NULL,
302
    meadow_type             VARCHAR(255) NOT NULL,
303 304
    submission_cmd_args     VARCHAR(255) NOT NULL DEFAULT '',
    worker_cmd_args         VARCHAR(255) NOT NULL DEFAULT '',
305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325

    PRIMARY KEY(resource_class_id, meadow_type)
);


/**
@header Job-related
@colour #1D73DA
*/

/**
@table  job

@colour #1D73DA

@desc The job is the heart of this system.  It is the kiosk or blackboard
    where workers find things to do and then post work for other works to do.
    These jobs are created prior to work being done, are claimed by workers,
    are updated as the work is done, with a final update on completion.

@column job_id                  autoincrement id
326
@column prev_job_id             previous job which created this one
327 328
@column analysis_id             the analysis_id needed to accomplish this job.
@column input_id                input data passed into Analysis:RunnableDB to control the work
329 330
@column param_id_stack          a CSV of job_ids whose input_ids contribute to the stack of local variables for the job
@column accu_id_stack           a CSV of job_ids whose accu's contribute to the stack of local variables for the job
331
@column role_id                 links to the Role that claimed this job (NULL means it has never been claimed)
332 333
@column status                  state the job is in
@column retry_count             number times job had to be reset when worker failed to run it
334
@column when_completed          when the job was completed
335 336 337 338 339
@column runtime_msec            how long did it take to execute the job (or until the moment it failed)
@column query_count             how many SQL queries were run during this job
@column semaphore_count         if this count is >0, the job is conditionally blocked (until this count drops to 0 or below). Default=0 means "nothing is blocking me by default".
@column semaphored_job_id       the job_id of job S that is waiting for this job to decrease S's semaphore_count. Default=NULL means "I'm not blocking anything by default".
*/
340 341

CREATE TABLE job (
Leo Gordon's avatar
Leo Gordon committed
342
    job_id                  INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
343
    prev_job_id             INTEGER              DEFAULT NULL,      -- the job that created this one using a dataflow rule
Leo Gordon's avatar
Leo Gordon committed
344 345
    analysis_id             INTEGER     NOT NULL,
    input_id                CHAR(255)   NOT NULL,
346 347
    param_id_stack          CHAR(64)    NOT NULL DEFAULT '',
    accu_id_stack           CHAR(64)    NOT NULL DEFAULT '',
348
    role_id                 INTEGER              DEFAULT NULL,
349
    status                  TEXT        NOT NULL DEFAULT 'READY',   -- expected values: 'SEMAPHORED','READY','CLAIMED','COMPILATION','PRE_CLEANUP','FETCH_INPUT','RUN','WRITE_OUTPUT','POST_HEALTHCHECK','POST_CLEANUP','DONE','FAILED','PASSED_ON'
Leo Gordon's avatar
Leo Gordon committed
350
    retry_count             INTEGER     NOT NULL DEFAULT 0,
351
    when_completed          TIMESTAMP            DEFAULT NULL,
Leo Gordon's avatar
Leo Gordon committed
352 353 354 355 356
    runtime_msec            INTEGER              DEFAULT NULL,
    query_count             INTEGER              DEFAULT NULL,

    semaphore_count         INTEGER     NOT NULL DEFAULT 0,
    semaphored_job_id       INTEGER              DEFAULT NULL
357
);
358
CREATE UNIQUE INDEX job_input_id_stacks_analysis_idx ON job (input_id, param_id_stack, accu_id_stack, analysis_id);
Leo Gordon's avatar
Leo Gordon committed
359
CREATE        INDEX job_analysis_status_retry_idx ON job (analysis_id, status, retry_count);
360
CREATE        INDEX job_role_id_status_id_idx ON job (role_id);
361 362


363 364
/**
@table  job_file
365

366
@colour #1D73DA
367

368 369 370 371
@desc   For testing/debugging purposes both STDOUT and STDERR streams of each Job
        can be redirected into a separate log file.
        This table holds filesystem paths to one or both of those files.
        There is max one entry per job_id and retry.
372

373 374
@column job_id             foreign key
@column retry              copy of retry_count of job as it was run
375
@column role_id            links to the Role that claimed this job
376 377 378
@column stdout_file        path to the job's STDOUT log
@column stderr_file        path to the job's STDERR log
*/
379 380

CREATE TABLE job_file (
Leo Gordon's avatar
Leo Gordon committed
381 382
    job_id                  INTEGER     NOT NULL,
    retry                   INTEGER     NOT NULL,
383
    role_id                 INTEGER     NOT NULL,
Leo Gordon's avatar
Leo Gordon committed
384 385
    stdout_file             VARCHAR(255),
    stderr_file             VARCHAR(255),
386

Leo Gordon's avatar
Leo Gordon committed
387
    PRIMARY KEY (job_id, retry)
388
);
389
CREATE        INDEX job_file_role_id_idx ON job_file (role_id);
390 391


392 393
/**
@table  accu
394

395 396 397
@colour #1D73DA

@desc   Accumulator for funneled dataflow.
398

399 400 401 402 403 404
@column sending_job_id     semaphoring job in the "box"
@column receiving_job_id   semaphored job outside the "box"
@column struct_name        name of the structured parameter
@column key_signature      locates the part of the structured parameter
@column value              value of the part
*/
405

406
CREATE TABLE accu (
Leo Gordon's avatar
Leo Gordon committed
407 408 409 410
    sending_job_id          INTEGER,
    receiving_job_id        INTEGER     NOT NULL,
    struct_name             VARCHAR(255) NOT NULL,
    key_signature           VARCHAR(255) NOT NULL,
411
    value                   TEXT
412
);
Leo Gordon's avatar
Leo Gordon committed
413 414
CREATE        INDEX accu_sending_idx   ON accu (sending_job_id);
CREATE        INDEX accu_receiving_idx ON accu (receiving_job_id);
415 416


417 418
/**
@table  analysis_data
419

420 421 422 423 424 425 426 427
@colour #1D73DA

@desc   A generic blob-storage hash.
        Currently the only legitimate use of this table is "overflow" of job.input_ids:
        when they grow longer than 254 characters the real data is stored in analysis_data instead,
        and the input_id contains the corresponding analysis_data_id.

@column analysis_data_id    primary id
428
@column md5sum              checksum over the data to quickly detect (potential) collisions
429 430 431 432
@column data                text blob which holds the data
*/

CREATE TABLE analysis_data (
Leo Gordon's avatar
Leo Gordon committed
433
    analysis_data_id        INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
434
    md5sum                  CHAR(32)    NOT NULL,
435
    data                    TEXT        NOT NULL
436
);
437
CREATE        INDEX analysis_data_idx ON analysis_data (md5sum);
438 439


440
/**
441
@header execution tables
442 443 444 445 446 447 448 449 450
@colour #24DA06
*/

/**
@table  worker

@colour #24DA06

@desc Entries of this table correspond to Worker objects of the API.
Leo Gordon's avatar
Leo Gordon committed
451 452 453 454 455 456
        Workers are created by inserting into this table
        so that there is only one instance of a Worker object in the database.
        As Workers live and do work, they update this table, and when they die they update again.

@column worker_id           unique ID of the Worker
@column meadow_type         type of the Meadow it is running on
457 458 459
@column meadow_name         name of the Meadow it is running on (for meadow_type=='LOCAL' it is the same as meadow_host)
@column meadow_host         execution host name
@column meadow_user         scheduling/execution user name (within the Meadow)
Leo Gordon's avatar
Leo Gordon committed
460 461 462 463
@column process_id          identifies the Worker process on the Meadow (for 'LOCAL' is the OS PID)
@column resource_class_id   links to Worker's resource class
@column work_done           how many jobs the Worker has completed successfully
@column status              current status of the Worker
464 465 466 467
@column when_born           when the Worker process was started
@column when_checked_in     when the Worker last checked into the database
@column when_seen           when the Worker was last seen by the Meadow
@column when_died           if defined, when the Worker died (or its premature death was first detected by GC)
Leo Gordon's avatar
Leo Gordon committed
468 469
@column cause_of_death      if defined, why did the Worker exit (or why it was killed)
@column log_dir             if defined, a filesystem directory where this Worker's output is logged
470
*/
471

472
CREATE TABLE worker (
Leo Gordon's avatar
Leo Gordon committed
473
    worker_id               INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
474 475
    meadow_type             VARCHAR(255) NOT NULL,
    meadow_name             VARCHAR(255) NOT NULL,
476 477
    meadow_host             VARCHAR(255) NOT NULL,
    meadow_user             VARCHAR(255)         DEFAULT NULL,
478
    process_id              VARCHAR(255) NOT NULL,
Leo Gordon's avatar
Leo Gordon committed
479
    resource_class_id       INTEGER              DEFAULT NULL,
480 481 482
    work_done               INTEGER      NOT NULL DEFAULT 0,
    status                  VARCHAR(255) NOT NULL DEFAULT 'READY', /* enum('SPECIALIZATION','COMPILATION','READY','JOB_LIFECYCLE','DEAD') DEFAULT 'READY' NOT NULL */
    when_born               TIMESTAMP    NOT NULL DEFAULT CURRENT_TIMESTAMP,
483 484 485
    when_checked_in         TIMESTAMP            DEFAULT NULL,
    when_seen               TIMESTAMP            DEFAULT NULL,
    when_died               TIMESTAMP            DEFAULT NULL,
486
    cause_of_death          TEXT                 DEFAULT NULL, /* enum('NO_ROLE', 'NO_WORK', 'JOB_LIMIT', 'HIVE_OVERLOAD', 'LIFESPAN', 'CONTAMINATED', 'RELOCATED', 'KILLED_BY_USER', 'MEMLIMIT', 'RUNLIMIT', 'SEE_MSG', 'UNKNOWN') DEFAULT NULL */
Leo Gordon's avatar
Leo Gordon committed
487
    log_dir                 VARCHAR(255)         DEFAULT NULL
488
);
489
CREATE INDEX worker_meadow_type_meadow_name_process_id_idx ON worker (meadow_type, meadow_name, process_id);
490 491


492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523
/**
@table  role

@colour #24DA06

@desc Entries of this table correspond to Role objects of the API.
        When a Worker specializes, it acquires a Role,
        which is a temporary link between the Worker and a resource-compatible Analysis.

@column role_id             unique ID of the Role
@column worker_id           the specialized Worker
@column analysis_id         the Analysis into which the Worker specialized
@column when_started        when this Role started
@column when_finished       when this Role finished. NULL may either indicate it is still running or was killed by an external force.
@column attempted_jobs      counter of the number of attempts
@column done_jobs           counter of the number of successful attempts
*/

CREATE TABLE role (
    role_id                 INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
    worker_id               INTEGER     NOT NULL,
    analysis_id             INTEGER     NOT NULL,
    when_started            TIMESTAMP   NOT NULL DEFAULT CURRENT_TIMESTAMP,
    when_finished           TIMESTAMP            DEFAULT NULL,
    attempted_jobs          INTEGER     NOT NULL DEFAULT 0,
    done_jobs               INTEGER     NOT NULL DEFAULT 0
);
CREATE        INDEX role_worker_id_idx ON role (worker_id);
CREATE        INDEX role_analysis_id_idx ON role (analysis_id);



524 525 526 527
/**
@header Logging and monitoring
@colour #F4D20C
*/
528

529 530 531 532 533 534
/**
@table  worker_resource_usage

@colour #F4D20C

@desc   A table with post-mortem resource usage statistics of a Worker.
535 536 537
	This table is not automatically populated: you first need to run
	load_resource_usage.pl. Note that some meadows (like LOCAL) do not
	support post-mortem inspection of resource usage
538 539 540 541 542 543

@column          worker_id  links to the worker table
@column        exit_status  meadow-dependent, in case of LSF it's usually 'done' (normal) or 'exit' (abnormal)
@column           mem_megs  how much memory the Worker process used
@column          swap_megs  how much swap the Worker process used
@column        pending_sec  time spent by the process in the queue before it became a Worker
544 545
@column            cpu_sec  cpu time (in seconds) used by the Worker process. It is often lower than the walltime because of time spent in I/O waits, but it can also be higher if the process is multi-threaded
@column       lifespan_sec  walltime (in seconds) used by the Worker process. It is often higher than the sum of its jobs' "runtime_msec" because of the overhead from the Worker itself
546 547 548 549 550 551 552 553
@column   exception_status  meadow-specific flags, in case of LSF it can be 'underrun', 'overrun' or 'idle'
*/

CREATE TABLE worker_resource_usage (
    worker_id               INTEGER         NOT NULL,
    exit_status             VARCHAR(255)    DEFAULT NULL,
    mem_megs                FLOAT           DEFAULT NULL,
    swap_megs               FLOAT           DEFAULT NULL,
554
    pending_sec             FLOAT           DEFAULT NULL,
555
    cpu_sec                 FLOAT           DEFAULT NULL,
556
    lifespan_sec            FLOAT           DEFAULT NULL,
557 558 559 560 561 562
    exception_status        VARCHAR(255)    DEFAULT NULL,

    PRIMARY KEY (worker_id)
);


563 564
/**
@table  log_message
565

566
@colour #F4D20C
567 568 569 570 571 572 573 574

@desc   When a Job or a job-less Worker (job_id=NULL) throws a "die" message
        for any reason, the message is recorded in this table.
        It may or may not indicate that the job was unsuccessful via is_error flag.
        Also $self->warning("...") messages are recorded with is_error=0.

@column log_message_id  an autoincremented primary id of the message
@column         job_id  the id of the job that threw the message (or NULL if it was outside of a message)
575
@column        role_id  the 'current' role
576
@column      worker_id  the 'current' worker
577
@column    when_logged  when the message was thrown
578 579 580 581 582 583 584
@column          retry  retry_count of the job when the message was thrown (or NULL if no job)
@column         status  of the job or worker when the message was thrown
@column            msg  string that contains the message
@column       is_error  binary flag
*/

CREATE TABLE log_message (
Leo Gordon's avatar
Leo Gordon committed
585 586
    log_message_id          INTEGER     NOT NULL PRIMARY KEY AUTOINCREMENT,
    job_id                  INTEGER              DEFAULT NULL,
587
    role_id                 INTEGER              DEFAULT NULL,
588
    worker_id               INTEGER              DEFAULT NULL,
589
    when_logged             TIMESTAMP   NOT NULL DEFAULT CURRENT_TIMESTAMP,
Leo Gordon's avatar
Leo Gordon committed
590
    retry                   INTEGER              DEFAULT NULL,
591
    status                  TEXT        NOT NULL DEFAULT 'UNKNOWN', /* enum('UNKNOWN','SPECIALIZATION','COMPILATION','CLAIMED','READY','PRE_CLEANUP','FETCH_INPUT','RUN','WRITE_OUTPUT','POST_CLEANUP','PASSED_ON') DEFAULT 'UNKNOWN' */
Leo Gordon's avatar
Leo Gordon committed
592 593
    msg                     TEXT,
    is_error                SMALLINT
594
);
Leo Gordon's avatar
Leo Gordon committed
595 596
CREATE        INDEX log_message_worker_idx ON log_message (worker_id);
CREATE        INDEX log_message_job_idx ON log_message (job_id);
597 598 599 600 601 602


/**
@table  analysis_stats_monitor

@colour #F4D20C
603

604 605
@desc   A regular timestamped snapshot of the analysis_stats table.

606
@column when_logged             when this snapshot was taken
607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629

@column analysis_id             foreign-keyed to the corresponding analysis_base entry
@column batch_size              how many jobs are claimed in one claiming operation before Worker starts executing them
@column hive_capacity           a reciprocal limiter on the number of Workers running at the same time (dependent on Workers of other Analyses)
@column status                  cached state of the Analysis

@column total_job_count         total number of Jobs of this Analysis
@column semaphored_job_count    number of Jobs of this Analysis that are in SEMAPHORED state
@column ready_job_count         number of Jobs of this Analysis that are in READY state
@column done_job_count          number of Jobs of this Analysis that are in DONE state
@column failed_job_count        number of Jobs of this Analysis that are in FAILED state

@column num_running_workers     number of running Workers of this Analysis

@column behaviour               whether hive_capacity is set or is dynamically calculated based on timers
@column input_capacity          used to compute hive_capacity in DYNAMIC mode
@column output_capacity         used to compute hive_capacity in DYNAMIC mode

@column avg_msec_per_job        weighted average used to compute DYNAMIC hive_capacity
@column avg_input_msec_per_job  weighted average used to compute DYNAMIC hive_capacity
@column avg_run_msec_per_job    weighted average used to compute DYNAMIC hive_capacity
@column avg_output_msec_per_job weighted average used to compute DYNAMIC hive_capacity

630
@column when_updated            when this entry was last updated
631 632
@column sync_lock               a binary lock flag to prevent simultaneous updates
*/
633 634

CREATE TABLE analysis_stats_monitor (
635
    when_logged             TIMESTAMP   NOT NULL DEFAULT CURRENT_TIMESTAMP,
Leo Gordon's avatar
Leo Gordon committed
636 637 638 639 640 641 642 643 644 645 646

    analysis_id             INTEGER     NOT NULL,
    batch_size              INTEGER     NOT NULL DEFAULT 1,
    hive_capacity           INTEGER              DEFAULT NULL,
    status                  TEXT        NOT NULL DEFAULT 'EMPTY', /* enum('BLOCKED', 'LOADING', 'SYNCHING', 'EMPTY', 'READY', 'WORKING', 'ALL_CLAIMED', 'DONE', 'FAILED') DEFAULT 'EMPTY' NOT NULL, */

    total_job_count         INTEGER     NOT NULL DEFAULT 0,
    semaphored_job_count    INTEGER     NOT NULL DEFAULT 0,
    ready_job_count         INTEGER     NOT NULL DEFAULT 0,
    done_job_count          INTEGER     NOT NULL DEFAULT 0,
    failed_job_count        INTEGER     NOT NULL DEFAULT 0,
647

Leo Gordon's avatar
Leo Gordon committed
648 649 650 651 652 653 654 655 656 657 658
    num_running_workers     INTEGER     NOT NULL DEFAULT 0,

    behaviour               TEXT        NOT NULL DEFAULT 'STATIC', /* enum('STATIC', 'DYNAMIC') DEFAULT 'STATIC' NOT NULL, */
    input_capacity          INTEGER     NOT NULL DEFAULT 4,
    output_capacity         INTEGER     NOT NULL DEFAULT 4,

    avg_msec_per_job        INTEGER              DEFAULT NULL,
    avg_input_msec_per_job  INTEGER              DEFAULT NULL,
    avg_run_msec_per_job    INTEGER              DEFAULT NULL,
    avg_output_msec_per_job INTEGER              DEFAULT NULL,

659
    when_updated            TIMESTAMP            DEFAULT NULL,
Leo Gordon's avatar
Leo Gordon committed
660
    sync_lock               SMALLINT    NOT NULL DEFAULT 0
661 662
);