tables.sql 12 KB
Newer Older
Jessica Severin's avatar
Jessica Severin committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
------------------------------------------------------------------------------------
--
-- Table structure for table 'hive'
--
-- overview:
--   Table which tracks the workers of a hive as they exist out in the world.
--   Workers are created by inserting into this table so that there is only every
--   one instance of a worker object in the world.  As workers live and do work,
--   they update this table, and when they die they update.
--
-- semantics:
--

CREATE TABLE hive (
  hive_id          int(10) NOT NULL auto_increment,
  analysis_id      int(10) NOT NULL,
Abel Ureta-Vidal's avatar
Abel Ureta-Vidal committed
17
  beekeeper        varchar(80) DEFAULT '' NOT NULL,
Jessica Severin's avatar
Jessica Severin committed
18
  host	           varchar(40) DEFAULT '' NOT NULL,
19
  process_id       varchar(40) DEFAULT '' NOT NULL,
Jessica Severin's avatar
Jessica Severin committed
20
  work_done        int(11) DEFAULT '0' NOT NULL,
21
  status           enum('READY','GET_INPUT','RUN','WRITE_OUTPUT','DEAD') DEFAULT 'READY' NOT NULL,
Jessica Severin's avatar
Jessica Severin committed
22 23 24
  born	           datetime NOT NULL,
  last_check_in    datetime NOT NULL,
  died             datetime DEFAULT NULL,
25 26
  cause_of_death   enum('', 'NO_WORK', 'JOB_LIMIT', 'HIVE_OVERLOAD', 'LIFESPAN', 'FATALITY') DEFAULT '' NOT NULL,
  PRIMARY KEY (hive_id),
27
  INDEX analysis_status (analysis_id, status)
Jessica Severin's avatar
Jessica Severin committed
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
);


------------------------------------------------------------------------------------
--
-- Table structure for table 'dataflow_rule'
--
-- overview:
--   Extension of simple_rule design except that goal(to) is now in extended URL format e.g.
--   mysql://ensadmin:<pass>@ecs2:3361/compara_hive_test?analysis.logic_name='blast_NCBI34'
--   (full network address of an analysis).  The only requirement is that there are rows in 
--   the analysis_job, analysis, dataflow_rule, and hive tables so that the following join
--   works on the same database 
--   WHERE analysis.analysis_id = dataflow_rule.from_analysis_id 
--   AND   analysis.analysis_id = analysis_job.analysis_id
--   AND   analysis.analysis_id = hive.analysis_id
--
--   These are the rules used to create entries in the analysis_job table where the
--   input_id (control data) is passed from one analysis to the next to define work.
--  
--   The analysis table will be extended so that it can specify different read and write
--   databases, with the default being the database the analysis is on
--
-- semantics:
52
--   dataflow_rule_id     - internal ID
Jessica Severin's avatar
Jessica Severin committed
53 54 55 56 57
--   from_analysis_id     - foreign key to analysis table analysis_id
--   to_analysis_url      - foreign key to net distributed analysis logic_name reference
--   branch_code          - joined to analysis_job.branch_code to allow branching

CREATE TABLE dataflow_rule (
58
  dataflow_rule_id    int(10) unsigned not null auto_increment,
Jessica Severin's avatar
Jessica Severin committed
59 60 61 62
  from_analysis_id    int(10) unsigned NOT NULL,
  to_analysis_url     varchar(255) default '' NOT NULL,
  branch_code         int(10) default 1 NOT NULL,

63
  PRIMARY KEY (dataflow_rule_id),
Jessica Severin's avatar
Jessica Severin committed
64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86
  UNIQUE (from_analysis_id, to_analysis_url)
);


------------------------------------------------------------------------------------
--
-- Table structure for table 'analysis_ctrl_rule'
--
-- overview:
--   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 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.
--
--   This process of watching conditions and flipping the ctrled_analysis state
--   will be accomplished by another automous agent (CtrlWatcher.pm)
--
-- semantics:
--   condition_analysis_url  - foreign key to net distributed analysis reference
--   ctrled_analysis_id      - foreign key to analysis table analysis_id

87
CREATE TABLE analysis_ctrl_rule (
Jessica Severin's avatar
Jessica Severin committed
88 89 90
  condition_analysis_url     varchar(255) default '' NOT NULL,
  ctrled_analysis_id         int(10) unsigned NOT NULL,

91
  UNIQUE (condition_analysis_url, ctrled_analysis_id)
Jessica Severin's avatar
Jessica Severin committed
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106
);


------------------------------------------------------------------------------------
--
-- Table structure for table 'analysis_job'
--
-- overview:
--   The analysis_job is the heart of this sytem.  It is the kiosk or blackboard
--   where workers find things to do and then post work for other works to do.
--   The job_claim is a UUID set with an UPDATE LIMIT by worker as they fight
--   over the work.  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.
--
-- semantics:
107 108 109
--   analysis_job_id         - autoincrement id
--   prev_analysis_job_id    - previous analysis_job which created this one (and passed input_id)
--   analysis_id             - the analysis_id needed to accomplish this job.
110
--   input_id                - input data passed into Analysis:RunnableDB to control the work
111 112 113 114 115 116
--   job_claim               - UUID set by workers as the fight over jobs
--   hive_id                 - link to hive table to define which worker claimed this job
--   status                  - state the job is in
--   retry_count             - number times job had to be reset when worker failed to run it
--   completed               - timestamp when job was completed
--   branch_code             - switch-like branching control, default=1 (ie true)
Jessica Severin's avatar
Jessica Severin committed
117 118

CREATE TABLE analysis_job (
119 120 121
  analysis_job_id           int(10) NOT NULL auto_increment,
  prev_analysis_job_id      int(10) NOT NULL,  #analysis_job which created this from rules
  analysis_id               int(10) NOT NULL,
122 123
  input_id                  char(255) not null,
  job_claim                 char(40) NOT NULL default '', #UUID
124 125 126 127 128
  hive_id                   int(10) NOT NULL,
  status                    enum('READY','BLOCKED','CLAIMED','GET_INPUT','RUN','WRITE_OUTPUT','DONE','FAILED') DEFAULT 'READY' NOT NULL,
  retry_count               int(10) default 0 not NULL,
  completed                 datetime NOT NULL,
  branch_code               int(10) default 1 NOT NULL,
129 130
  runtime_msec              int(10) default 0 NOT NULL, 
  query_count               int(10) default 0 NOT NULL, 
Jessica Severin's avatar
Jessica Severin committed
131 132

  PRIMARY KEY                  (analysis_job_id),
133
  UNIQUE KEY input_id_analysis (input_id, analysis_id),
134 135
  INDEX claim_analysis_status  (job_claim, analysis_id, status),
  INDEX analysis_status        (analysis_id, status),
136
  INDEX hive_id                (hive_id)
Jessica Severin's avatar
Jessica Severin committed
137 138 139 140 141 142 143 144 145 146 147 148 149 150 151
);


------------------------------------------------------------------------------------
--
-- Table structure for table 'analysis_job_file'
--
-- overview:
--   Table which holds paths to files created by an analysis_job
--   e.g. STDOUT STDERR, temp directory
--   or output data files created by the RunnableDB
--   There can only be one entry of a certain type for a given analysis_job
--
-- semantics:
--   analysis_job_id    - foreign key
152 153
--   hive_id            - link to hive table to define which worker claimed this job
--   retry              - copy of retry_count of job as it was run
Jessica Severin's avatar
Jessica Severin committed
154 155 156 157 158
--   type               - type of file e.g. STDOUT, STDERR, TMPDIR, ...
--   path               - path to file or directory

CREATE TABLE analysis_job_file (
  analysis_job_id         int(10) NOT NULL,
159 160
  hive_id                 int(10) NOT NULL,
  retry                   int(10) NOT NULL,
Jessica Severin's avatar
Jessica Severin committed
161 162 163
  type                    varchar(16) NOT NULL default '',
  path                    varchar(255) NOT NULL,
  
164 165
  UNIQUE KEY job_hive_type  (analysis_job_id, hive_id, type),
  INDEX hive_id             (hive_id)
Jessica Severin's avatar
Jessica Severin committed
166 167 168
);


169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186
------------------------------------------------------------------------------------
--
-- Table structure for table 'analysis_data'
--
-- overview:
--   Table which holds LONGTEXT data for use by the analysis system.
--   This data is general purpose and it's up to each analysis to
--   determine how to use it
--
-- semantics:
--   analysis_data_id   - primary id
--   data               - text blob which holds the data

CREATE TABLE analysis_data (
  analysis_data_id  int(10) NOT NULL auto_increment,
  data              longtext,

  PRIMARY KEY (analysis_data_id),
187
  KEY data (data(100))
188 189 190
);


Jessica Severin's avatar
Jessica Severin committed
191 192 193 194 195 196 197 198 199 200 201
------------------------------------------------------------------------------------
--
-- Table structure for table 'analysis_stats'
--
-- overview:
--   Parallel table to analysis 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
--
-- semantics:
202 203 204
--   analysis_id          - foreign key to analysis table
--   status               - overview status of the analysis_jobs (cached state)
--   failed_job_tolerance - % of tolerated failed jobs
Jessica Severin's avatar
Jessica Severin committed
205 206 207

CREATE TABLE analysis_stats (
  analysis_id           int(10) NOT NULL,
208
  status                enum('BLOCKED', 'LOADING', 'SYNCHING', 'READY', 'WORKING', 'ALL_CLAIMED', 'DONE', 'FAILED')
209
                          DEFAULT 'READY' NOT NULL,
210
  batch_size            int(10) default 1 NOT NULL,
211 212 213 214
  avg_msec_per_job      int(10) default 0 NOT NULL,
  avg_input_msec_per_job      int(10) default 0 NOT NULL,
  avg_run_msec_per_job      int(10) default 0 NOT NULL,
  avg_output_msec_per_job      int(10) default 0 NOT NULL,
215
  hive_capacity         int(10) default 1 NOT NULL,
216 217 218
  behaviour             enum('STATIC', 'DYNAMIC') DEFAULT 'STATIC' NOT NULL,
  input_capacity        int(10) default 4 NOT NULL,
  output_capacity       int(10) default 4 NOT NULL,
219 220 221
  total_job_count       int(10) NOT NULL,
  unclaimed_job_count   int(10) NOT NULL,
  done_job_count        int(10) NOT NULL,
222
  max_retry_count       int(10) default 3 NOT NULL,
223
  failed_job_count      int(10) NOT NULL,
224
  failed_job_tolerance  int(10) default 0 NOT NULL,
225
  num_running_workers   int(10) default 0 NOT NULL,
Jessica Severin's avatar
Jessica Severin committed
226
  num_required_workers  int(10) NOT NULL,
Jessica Severin's avatar
Jessica Severin committed
227
  last_update           datetime NOT NULL,
228
  sync_lock             int(10) default 0 NOT NULL,
Jessica Severin's avatar
Jessica Severin committed
229 230 231 232
  
  UNIQUE KEY   (analysis_id)
);

233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
CREATE TABLE analysis_stats_monitor (
  time                  datetime NOT NULL default '0000-00-00 00:00:00',
  analysis_id           int(10) NOT NULL,
  status                enum('BLOCKED', 'LOADING', 'SYNCHING', 'READY', 'WORKING', 'ALL_CLAIMED', 'DONE', 'FAILED')
                          DEFAULT 'READY' NOT NULL,
  batch_size            int(10) default 1 NOT NULL,
  avg_msec_per_job      int(10) default 0 NOT NULL,
  avg_input_msec_per_job      int(10) default 0 NOT NULL,
  avg_run_msec_per_job      int(10) default 0 NOT NULL,
  avg_output_msec_per_job      int(10) default 0 NOT NULL,
  hive_capacity         int(10) default 1 NOT NULL,
  behaviour             enum('STATIC', 'DYNAMIC') DEFAULT 'STATIC' NOT NULL,
  input_capacity        int(10) default 4 NOT NULL,
  output_capacity       int(10) default 4 NOT NULL,
  total_job_count       int(10) NOT NULL,
  unclaimed_job_count   int(10) NOT NULL,
  done_job_count        int(10) NOT NULL,
  max_retry_count       int(10) default 3 NOT NULL,
  failed_job_count      int(10) NOT NULL,
  failed_job_tolerance  int(10) default 0 NOT NULL,
  num_running_workers   int(10) default 0 NOT NULL,
  num_required_workers  int(10) NOT NULL,
  last_update           datetime NOT NULL,
  sync_lock             int(10) default 0 NOT NULL
);

259 260 261 262 263 264 265 266 267 268 269 270 271 272
------------------------------------------------------------------------------------
--
-- Table structure for table 'monitor'
--
-- overview:
--   This table stores information about hive performance.
--
-- semantics:
--   time           - timestamp
--   workers        - number of running workers
--   throughput     - average numb of completed jobs per sec. of the hive
--                    (this number is calculated using running workers only)
--   per_worker     - average numb of completed jobs per sec. per worker
--                    (this number is calculated using running workers only)
273
--   analysis       - analysis(es) running at that time
274 275 276 277 278 279

CREATE TABLE monitor (
  time                  datetime NOT NULL default '0000-00-00 00:00:00',
  workers               int(10) NOT NULL default '0',
  throughput            float default NULL,
  per_worker            float default NULL,
280
  analysis              varchar(255) default NULL
281 282
);