Megatest

Check-in [93bcd23d16]
Login
Overview
Comment:Job queue for adjutant
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | v1.65-adjutant
Files: files | file ages | folders
SHA1: 93bcd23d16c6f8b0929b4dadceaabad9cb68eeae
User & Date: matt on 2020-09-15 22:44:27
Other Links: branch diff | manifest | tags
Context
2020-09-16
01:20
Almost sorta working check-in: b08c20c7c7 user: matt tags: v1.65-adjutant
2020-09-15
22:44
Job queue for adjutant check-in: 93bcd23d16 user: matt tags: v1.65-adjutant
2020-09-07
21:02
merged with v1.65-cleanup check-in: e373dd2861 user: mmgraham tags: v1.6568, v1.65-experiment
Changes

Modified api.scm from [4fa67bb6bd] to [0134572f5d].

249
250
251
252
253
254
255



256
257
258
259
260
261
262
263
264
265
266
267
268
269
                   ((tasks-get-last)            (apply tasks:get-last dbstruct params))

		   ;; NO SYNC DB
		   ((no-sync-set)               (apply db:no-sync-set         *no-sync-db* params))
		   ((no-sync-get/default)       (apply db:no-sync-get/default *no-sync-db* params))
		   ((no-sync-del!)              (apply db:no-sync-del!        *no-sync-db* params))
		   ((no-sync-get-lock)          (apply db:no-sync-get-lock    *no-sync-db* params))



		   
                   ;; ARCHIVES
                   ;; ((archive-get-allocations)   
                   ((archive-register-disk)     (apply db:archive-register-disk dbstruct params))
                   ((archive-register-block-name)(apply db:archive-register-block-name dbstruct params))
                   ;; ((archive-allocate-testsuite/area-to-block)(apply db:archive-allocate-testsuite/area-to-block dbstruct block-id testsuite-name areakey))

                   ;;======================================================================
                   ;; READ ONLY QUERIES
                   ;;======================================================================

                   ;; KEYS
                   ((get-key-val-pairs)               (apply db:get-key-val-pairs dbstruct params))
                   ((get-keys)                        (db:get-keys dbstruct))







>
>
>






|







249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
                   ((tasks-get-last)            (apply tasks:get-last dbstruct params))

		   ;; NO SYNC DB
		   ((no-sync-set)               (apply db:no-sync-set         *no-sync-db* params))
		   ((no-sync-get/default)       (apply db:no-sync-get/default *no-sync-db* params))
		   ((no-sync-del!)              (apply db:no-sync-del!        *no-sync-db* params))
		   ((no-sync-get-lock)          (apply db:no-sync-get-lock    *no-sync-db* params))
		   ((no-sync-add-job)           (apply db:no-sync-add-job     *no-sync-db* params))
		   ((no-sync-take-job)          (apply db:no-sync-take-job    *no-sync-db* params))
		   ((no-sync-job-records-clean) (apply db:no-sync-job-records-clean *no-sync-db* params))
		   
                   ;; ARCHIVES
                   ;; ((archive-get-allocations)   
                   ((archive-register-disk)     (apply db:archive-register-disk dbstruct params))
                   ((archive-register-block-name)(apply db:archive-register-block-name dbstruct params))
                   ;; ((archive-allocate-testsuite/area-to-block)(apply db:archive-allocate-testsuite/area-to-block dbstruct block-id testsuite-name areakey))
		   
                   ;;======================================================================
                   ;; READ ONLY QUERIES
                   ;;======================================================================

                   ;; KEYS
                   ((get-key-val-pairs)               (apply db:get-key-val-pairs dbstruct params))
                   ((get-keys)                        (db:get-keys dbstruct))

Modified common.scm from [e271de7a8e] to [085ec1a000].

3169
3170
3171
3172
3173
3174
3175


3176
3177
3178


3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194

3195


3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210

3211
3212
3213
3214
3215
3216
3217
3218
3219

3220
3221
3222
3223
3224
3225
3226
3227


3228
3229
3230
3231
3232
3233
3234
3235
;; 
;; [hosts]
;; arm cubie01 cubie02
;; x86_64 zeus xena myth01
;; allhosts #{g hosts arm} #{g hosts x86_64}
;; 
;; [host-types]


;; general #MTLOWESTLOAD #{g hosts allhosts}
;; arm     #MTLOWESTLOAD #{g hosts arm}
;; nbgeneral nbjob run JOBCOMMAND -log $MT_LINKTREE/$MT_TARGET/$MT_RUNNAME.$MT_TESTNAME-$MT_ITEM_PATH.lgo


;;
;; [host-rules]
;; # maxnload   => max normalized load
;; # maxnjobs   => max jobs per cpu
;; # maxjobrate => max jobs per second
;; general maxnload=1.1; maxnjobs=1.2; maxjobrate=0.1 
;; 
;; [launchers]
;; envsetup general
;; xor/%/n 4C16G
;; % nbgeneral
;; 
;; [jobtools]
;; # if defined and not "no" flexi-launcher will bypass "launcher" unless no match.
;; flexi-launcher yes  
;; launcher nbfake

;;


(define (common:get-launcher configdat testname itempath)
  (let ((fallback-launcher (configf:lookup configdat "jobtools" "launcher")))
    (if (and (configf:lookup configdat "jobtools" "flexi-launcher") ;; overrides launcher
	     (not (equal? (configf:lookup configdat "jobtools" "flexi-launcher") "no")))
	(let* ((launchers         (hash-table-ref/default configdat "launchers" '())))
	  (if (null? launchers)
	      fallback-launcher
	      (let loop ((hed (car launchers))
			 (tal (cdr launchers)))
		(let ((patt      (car hed))
		      (host-type (cadr hed)))
		  (if (tests:match patt testname itempath)
		      (begin
			(debug:print-info 2 *default-log-port* "Have flexi-launcher match for " testname "/" itempath " = " host-type)
			(let ((launcher (configf:lookup configdat "host-types" host-type)))

			  (if launcher
			      (let* ((launcher-parts (string-split launcher))
				     (launcher-exe   (car launcher-parts)))
				(if (equal? launcher-exe "#MTLOWESTLOAD") ;; this is our special case, we will find the lowest load and craft a nbfake commandline
				    (let host-loop ((targ-host (common:get-least-loaded-host (cdr launcher-parts) host-type configdat))
						    (count     100))
				      (if targ-host
					  (conc "remrun " targ-host)
					  (if (> count 0)

					      (begin
						(debug:print 0 *default-log-port* "INFO: Waiting for a host for host-type " host-type)
						(thread-sleep! (- 101 count))
						(host-loop (common:get-least-loaded-host (cdr launcher-parts) host-type configdat)
							   (- count 1)))
					      (begin
						(debug:print 0 *default-log-port* "FATAL: Failed to find a host from #MTLOWESTLOAD for host-type " host-type)
						(exit)))))


				    launcher))
			      (begin
				(debug:print-info 0 *default-log-port* "WARNING: no launcher found for host-type " host-type)
				(if (null? tal)
				    fallback-launcher
				    (loop (car tal)(cdr tal)))))))
		      ;; no match, try again
		      (if (null? tal)







>
>



>
>
















>

>
>
|










|


|
>









>








>
>
|







3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
;; 
;; [hosts]
;; arm cubie01 cubie02
;; x86_64 zeus xena myth01
;; allhosts #{g hosts arm} #{g hosts x86_64}
;; 
;; [host-types]
;; C/M/A      lets megatest know this launcher provides C cores, M bytes memory for architecture A 
;; 2/2G/arm   smart -cores 2 -memory 2G -arch arm
;; general #MTLOWESTLOAD #{g hosts allhosts}
;; arm     #MTLOWESTLOAD #{g hosts arm}
;; nbgeneral nbjob run JOBCOMMAND -log $MT_LINKTREE/$MT_TARGET/$MT_RUNNAME.$MT_TESTNAME-$MT_ITEM_PATH.lgo
;;
;; NOTE: host-rules is ONLY used for MTLOWESTLOAD
;;
;; [host-rules]
;; # maxnload   => max normalized load
;; # maxnjobs   => max jobs per cpu
;; # maxjobrate => max jobs per second
;; general maxnload=1.1; maxnjobs=1.2; maxjobrate=0.1 
;; 
;; [launchers]
;; envsetup general
;; xor/%/n 4C16G
;; % nbgeneral
;; 
;; [jobtools]
;; # if defined and not "no" flexi-launcher will bypass "launcher" unless no match.
;; flexi-launcher yes  
;; launcher nbfake
;; mode adjutant|normal (default is normal)
;;
;; mode is 'normal (i.e. directly use launcher) or 'adjutant (i.e. use adjutant)
;;
(define (common:get-launcher configdat testname itempath mode)
  (let ((fallback-launcher (configf:lookup configdat "jobtools" "launcher")))
    (if (and (configf:lookup configdat "jobtools" "flexi-launcher") ;; overrides launcher
	     (not (equal? (configf:lookup configdat "jobtools" "flexi-launcher") "no")))
	(let* ((launchers         (hash-table-ref/default configdat "launchers" '())))
	  (if (null? launchers)
	      fallback-launcher
	      (let loop ((hed (car launchers))
			 (tal (cdr launchers)))
		(let ((patt      (car hed))
		      (host-type (cadr hed)))
		  (if (tests:match patt testname itempath) ;; have a launcher match for this test
		      (begin
			(debug:print-info 2 *default-log-port* "Have flexi-launcher match for " testname "/" itempath " = " host-type)
			(let ((launcher (configf:lookup configdat "host-types" host-type))) ;; find the actual launcher from the host-types table
			  ;; if we are in adjutant mode then we want to return both host-type and launcher
			  (if launcher
			      (let* ((launcher-parts (string-split launcher))
				     (launcher-exe   (car launcher-parts)))
				(if (equal? launcher-exe "#MTLOWESTLOAD") ;; this is our special case, we will find the lowest load and craft a nbfake commandline
				    (let host-loop ((targ-host (common:get-least-loaded-host (cdr launcher-parts) host-type configdat))
						    (count     100))
				      (if targ-host
					  (conc "remrun " targ-host)
					  (if (> count 0)
					      
					      (begin
						(debug:print 0 *default-log-port* "INFO: Waiting for a host for host-type " host-type)
						(thread-sleep! (- 101 count))
						(host-loop (common:get-least-loaded-host (cdr launcher-parts) host-type configdat)
							   (- count 1)))
					      (begin
						(debug:print 0 *default-log-port* "FATAL: Failed to find a host from #MTLOWESTLOAD for host-type " host-type)
						(exit)))))
				    (case mode
				      ((adjutant) (list host-type launcher))
				      (else       launcher))))
			      (begin
				(debug:print-info 0 *default-log-port* "WARNING: no launcher found for host-type " host-type)
				(if (null? tal)
				    fallback-launcher
				    (loop (car tal)(cdr tal)))))))
		      ;; no match, try again
		      (if (null? tal)

Modified db.scm from [bbc6b8f83a] to [da01ec24cb].

2140
2141
2142
2143
2144
2145
2146

2147




2148
2149
2150











































2151
2152
2153
2154
2155
2156
2157
  (let* ((dbpath (db:dbfile-path))
	 (dbname (conc dbpath "/no-sync.db"))
	 (db-exists (common:file-exists? dbname))
	 (db     (sqlite3:open-database dbname)))
    (sqlite3:set-busy-handler! db (sqlite3:make-busy-timeout 136000))
    (if (not db-exists)
	(begin

	  (sqlite3:execute db "PRAGMA synchronous = 0;")




	  (sqlite3:execute db "CREATE TABLE IF NOT EXISTS no_sync_metadat (var TEXT,val TEXT, CONSTRAINT no_sync_metadat_constraint UNIQUE (var));")
	  (sqlite3:execute db "PRAGMA journal_mode=WAL;")))
    db))












































;; if we are not a server create a db handle. this is not finalized
;; so watch for problems. I'm still not clear if it is needed to manually
;; finalize sqlite3 dbs with the sqlite3 egg.
;;
(define (db:no-sync-db db-in)
  (mutex-lock! *db-access-mutex*)







>
|
>
>
>
>
|
|

>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>







2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
  (let* ((dbpath (db:dbfile-path))
	 (dbname (conc dbpath "/no-sync.db"))
	 (db-exists (common:file-exists? dbname))
	 (db     (sqlite3:open-database dbname)))
    (sqlite3:set-busy-handler! db (sqlite3:make-busy-timeout 136000))
    (if (not db-exists)
	(begin
	  (sqlite3:execute db "CREATE TABLE IF NOT EXISTS no_sync_metadat (var TEXT,val TEXT, CONSTRAINT no_sync_metadat_constraint UNIQUE (var));")
	  (sqlite3:execute db "PRAGMA journal_mode=WAL;")))
    ;; MOVE THIS TABLE CREATION TO THE (begin above in about six months (it is Sep 2020 right now).
    (sqlite3:execute db "CREATE TABLE IF NOT EXISTS jobs_queue (id INTEGER PRIMARY KEY, host_type TEXT, cores INTEGER, memory TEXT, vars TEXT, exekey TEXT, state TEXT, event_time INTEGER, last_update INTEGER);")
    ;; not sure I'll use this next one. I prefer if tests simply append to a file:
    ;;    last-update-seconds cpuload tmpspace rundirspace
    (sqlite3:execute db "CREATE TABLE IF NOT EXISTS test_extra_data (id INTEGER PRIMARY KEY, run_id INTEGER, test_id INTEGER, last_seen_running INTEGER);") 
    (sqlite3:execute db "PRAGMA synchronous = 0;")
    db))

(define (db:no-sync-add-job db-in host-type vars-list exekey)
  (sqlite3:execute (db:no-sync-db db-in) "INSERT INTO jobs_queue (host_type,vars,exekey,state,event_time,last_update) VALUES (?,?,?,?,?,?);"
		   host-type vars-list exekey "waiting" (current-seconds)(current-seconds)))

;; find next job (waiting longest) that matches host-type - future, we'll find jobs that fit if no exact match
(define (db:no-sync-take-job db-in host-type)
  (let* ((db   (db:no-sync-db db-in))
	 (stmt1 "SELECT id,host_type,vars,exekey,state,event_time,last_update FROM jobs_queue WHERE host_type=? AND state != 'taken' ORDER BY event_time ASC;")
	 (stmt1h (sqlite3:prepare db stmt1))
	 (stmt2  "UPDATE jobs_queue SET state='taken',last_update=? WHERE id=?;")
	 (stmt2h (sqlite3:prepare db stmt2))
	 (res    (sqlite3:with-transaction
		  db
		  (lambda ()
		    (let* ((matching-jobs (sqlite3:fold-row
					   (lambda (res . row) ;; id host-type vars exekey state event-time last-update)
					     (cons row res))
					   '()
					   stmt1h
					   host-type)))
		      (if (null? matching-jobs)
			  #f
			  (let ((choosen-one  (let loop ((tal matching-jobs)
							 (res #f)) ;; put bestest one in here
						(if (null? tal)
						    res
						    (let ((curr (car tal))
							  (rem  (cdr tal)))
						      curr) ;; here we will compare with res, if better candidate the loop with curr else loop with res
						    ))))
			    (if choosen-one ;; we need to mark it as taken
				(sqlite3:execute stmt2h (current-seconds) (car choosen-one)))
			    choosen-one)))))))
    (sqlite3:finalize! stmt1h) ;; it'd be nice to cache these and finalize on exit.
    (sqlite3:finalize! stmt2h)
    res))

;; clean out old jobs in queue, i.e. taken and event_time > 24 hrs ago
;;
(define (db:no-sync-job-records-clean db)
  (sqlite3:execute (db:no-sync-db db) "DELETE FROM jobs_queue WHERE state='taken' AND event_time < ?;" (- (current-seconds)(* 24 3600))))
		 

;; if we are not a server create a db handle. this is not finalized
;; so watch for problems. I'm still not clear if it is needed to manually
;; finalize sqlite3 dbs with the sqlite3 egg.
;;
(define (db:no-sync-db db-in)
  (mutex-lock! *db-access-mutex*)
2165
2166
2167
2168
2169
2170
2171
2172

2173
2174
2175
2176
2177
2178
2179
2180

(define (db:no-sync-set db var val)
  (sqlite3:execute (db:no-sync-db db) "INSERT OR REPLACE INTO no_sync_metadat (var,val) VALUES (?,?);" var val))

(define (db:no-sync-del! db var)
  (sqlite3:execute (db:no-sync-db db) "DELETE FROM no_sync_metadat WHERE var=?;" var))

(define (db:no-sync-get/default db var default)

  (let ((res default))
    (sqlite3:for-each-row
     (lambda (val)
       (set! res val))
     (db:no-sync-db db)
     "SELECT val FROM no_sync_metadat WHERE var=?;"
     var)
    (if res







|
>
|







2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229

(define (db:no-sync-set db var val)
  (sqlite3:execute (db:no-sync-db db) "INSERT OR REPLACE INTO no_sync_metadat (var,val) VALUES (?,?);" var val))

(define (db:no-sync-del! db var)
  (sqlite3:execute (db:no-sync-db db) "DELETE FROM no_sync_metadat WHERE var=?;" var))

(define (db:no-sync-get/default db-in var default)
  (let ((db  (db:no-sync-db db-in))
	(res default))
    (sqlite3:for-each-row
     (lambda (val)
       (set! res val))
     (db:no-sync-db db)
     "SELECT val FROM no_sync_metadat WHERE var=?;"
     var)
    (if res

Modified launch.scm from [9ff8002b65] to [d361edaf28].

1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386


1387
1388
1389
1390
1391
1392
1393
;; 3. create link from run dir to megatest runs area 
;; 4. remotely run the test on allocated host
;;    - could be ssh to host from hosts table (update regularly with load)
;;    - could be netbatch
;;      (launch-test db (cadr status) test-conf))
(define (launch-test test-id run-id run-info keyvals runname test-conf test-name test-path itemdat params)
  (mutex-lock! *launch-setup-mutex*) ;; setting variables and processing the testconfig is NOT thread-safe, reuse the launch-setup mutex
  (let* ( ;; (lock-key        (conc "test-" test-id))
	;; (got-lock        (let loop ((lock        (rmt:no-sync-get-lock lock-key))
	;; 			     (expire-time (+ (current-seconds) 15))) ;; give up on getting the lock and steal it after 15 seconds
	;; 		    (if (car lock)
	;; 			#t
	;; 			(if (> (current-seconds) expire-time)
	;; 			    (begin
	;; 			      (debug:print-info 0 *default-log-port* "Timed out waiting for a lock to launch test " keyvals " " runname " " test-name " " test-path)
	;; 			      (rmt:no-sync-del! lock-key) ;; destroy the lock
	;; 			      (loop (rmt:no-sync-get-lock lock-key) expire-time)) ;; 
	;; 			    (begin
	;; 			      (thread-sleep! 1)
	;; 			      (loop (rmt:no-sync-get-lock lock-key) expire-time))))))
	 (item-path       (item-list->path itemdat))
	 (contour         #f)) ;; NOT READY FOR THIS (args:get-arg "-contour")))


    (let loop ((delta        (- (current-seconds) *last-launch*))
	       (launch-delay (configf:lookup-number *configdat* "setup" "launch-delay" default: 0)))
      (if (> launch-delay delta)
	  (begin
	    (if (common:low-noise-print 1200 "test launch delay") ;; every two hours or so remind the user about launch delay.
		(debug:print-info 0 *default-log-port* "NOTE: test launches are delayed by " launch-delay " seconds. See megatest.config launch-delay setting to adjust.")) ;; launch of " test-name " for " (- launch-delay delta) " seconds"))
	    (thread-sleep! (- launch-delay delta))







|
<
<
<
<
<
<
<
<
<
<
<
<

|
>
>







1365
1366
1367
1368
1369
1370
1371
1372












1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
;; 3. create link from run dir to megatest runs area 
;; 4. remotely run the test on allocated host
;;    - could be ssh to host from hosts table (update regularly with load)
;;    - could be netbatch
;;      (launch-test db (cadr status) test-conf))
(define (launch-test test-id run-id run-info keyvals runname test-conf test-name test-path itemdat params)
  (mutex-lock! *launch-setup-mutex*) ;; setting variables and processing the testconfig is NOT thread-safe, reuse the launch-setup mutex
  (let* (;; locking code removed from here commented out and pasted at end of file












	 (item-path       (item-list->path itemdat))
	 (contour         #f)                         ;; NOT READY FOR THIS (args:get-arg "-contour")))
	 ;; launch-mode will be 'adjutant or 'normal
	 (launch-mode     (string->symbol (or (configf:lookup *configdat* "jobtools" "mode") "normal"))))
    (let loop ((delta        (- (current-seconds) *last-launch*))
	       (launch-delay (configf:lookup-number *configdat* "setup" "launch-delay" default: 0)))
      (if (> launch-delay delta)
	  (begin
	    (if (common:low-noise-print 1200 "test launch delay") ;; every two hours or so remind the user about launch delay.
		(debug:print-info 0 *default-log-port* "NOTE: test launches are delayed by " launch-delay " seconds. See megatest.config launch-delay setting to adjust.")) ;; launch of " test-name " for " (- launch-delay delta) " seconds"))
	    (thread-sleep! (- launch-delay delta))
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434

1435
1436
1437
1438
1439
1440
1441


1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
	   (subrun          (> (length (hash-table-ref/default tconfig "subrun"  '())) 0)) ;; send a flag to process a subrun
	   ;; (diskspace       (configf:lookup tconfig   "requirements" "diskspace"))
	   ;; (memory          (configf:lookup tconfig   "requirements" "memory"))
	   ;; (hosts           (configf:lookup *configdat* "jobtools"     "workhosts")) ;; I'm pretty sure this was never completed
	   (remote-megatest (configf:lookup *configdat* "setup" "executable"))
	   (run-time-limit  (or (configf:lookup  tconfig   "requirements" "runtimelim")
				(configf:lookup  *configdat* "setup" "runtimelim")))
	   ;; FIXME SOMEDAY: not good how this is so obtuse, this hack is to 
	   ;;                allow running from dashboard. Extract the path
	   ;;                from the called megatest and convert dashboard
	   ;;             	  or dboard to megatest
	   (local-megatest  (common:find-local-megatest))
	   #;(local-megatest  (let* ((lm  (car (argv)))
				   (dir (pathname-directory lm))
				   (exe (pathname-strip-directory lm)))

			      (conc (if dir (conc dir "/") "")
				    (case (string->symbol exe)
				      ((dboard)    "../megatest")
				      ((mtest)     "../megatest")
				      ((dashboard) "megatest")
				      (else exe)))))
	   (launcher        (common:get-launcher *configdat* test-name item-path)) ;; (configf:lookup *configdat* "jobtools"     "launcher"))


	   (test-sig        (conc (common:get-testsuite-name) ":" test-name ":" item-path)) ;; (item-list->path itemdat))) ;; test-path is the full path including the item-path
	   (work-area       #f)
	   (toptest-work-area #f) ;; for iterated tests the top test contains data relevant for all
	   (diskpath   #f)
	   (cmdparms   #f)
	   (fullcmd    #f) ;; (define a (with-output-to-string (lambda ()(write x))))
	   (mt-bindir-path #f)
	   (testinfo   (rmt:get-test-info-by-id run-id test-id))
	   (mt_target  (string-intersperse (map cadr keyvals) "/"))
	   (debug-param (append (if (args:get-arg "-debug")  (list "-debug" (args:get-arg "-debug")) '())
				(if (args:get-arg "-logging")(list "-logging") '())
				(if (configf:lookup *configdat* "misc" "profilesw")
				    (list (configf:lookup *configdat* "misc" "profilesw"))
				    '()))))
      ;; (if hosts (set! hosts (string-split hosts)))
      ;; set the megatest to be called on the remote host
      (if (not remote-megatest)(set! remote-megatest local-megatest)) ;; "megatest"))
      (set! mt-bindir-path (pathname-directory remote-megatest))
      (if launcher (set! launcher (string-split launcher)))
      ;; set up the run work area for this test
      (if (and (args:get-arg "-preclean") ;; user has requested to preclean for this run
	       (not (member (db:test-get-rundir testinfo)(list "n/a" "/tmp/badname")))) ;; n/a is a placeholder and thus not a read dir
	  (begin
	    (debug:print-info 0 *default-log-port* "attempting to preclean directory " (db:test-get-rundir testinfo) " for test " test-name "/" item-path)
	    (runs:remove-test-directory testinfo 'remove-data-only))) ;; remove data only, do not perturb the record
      







<
<
<
<

<
<
<
>
|
|
<
<
<
<
<
>
>
|

















|







1410
1411
1412
1413
1414
1415
1416




1417



1418
1419
1420





1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
	   (subrun          (> (length (hash-table-ref/default tconfig "subrun"  '())) 0)) ;; send a flag to process a subrun
	   ;; (diskspace       (configf:lookup tconfig   "requirements" "diskspace"))
	   ;; (memory          (configf:lookup tconfig   "requirements" "memory"))
	   ;; (hosts           (configf:lookup *configdat* "jobtools"     "workhosts")) ;; I'm pretty sure this was never completed
	   (remote-megatest (configf:lookup *configdat* "setup" "executable"))
	   (run-time-limit  (or (configf:lookup  tconfig   "requirements" "runtimelim")
				(configf:lookup  *configdat* "setup" "runtimelim")))




	   (local-megatest  (common:find-local-megatest))



	   (launcher        (let ((l (common:get-launcher *configdat* test-name item-path launcher-mode)))
			      (if (string? l)
				  (string-split l)





				  l))) ;; some nonhomogenuity here. '(cmd param1 param2 ...) OR '(host-type launcher)
	    ;; (item-list->path itemdat))) ;; test-path is the full path including the item-path
	   (test-sig        (conc (common:get-testsuite-name) ":" test-name ":" item-path))
	   (work-area       #f)
	   (toptest-work-area #f) ;; for iterated tests the top test contains data relevant for all
	   (diskpath   #f)
	   (cmdparms   #f)
	   (fullcmd    #f) ;; (define a (with-output-to-string (lambda ()(write x))))
	   (mt-bindir-path #f)
	   (testinfo   (rmt:get-test-info-by-id run-id test-id))
	   (mt_target  (string-intersperse (map cadr keyvals) "/"))
	   (debug-param (append (if (args:get-arg "-debug")  (list "-debug" (args:get-arg "-debug")) '())
				(if (args:get-arg "-logging")(list "-logging") '())
				(if (configf:lookup *configdat* "misc" "profilesw")
				    (list (configf:lookup *configdat* "misc" "profilesw"))
				    '()))))
      ;; (if hosts (set! hosts (string-split hosts)))
      ;; set the megatest to be called on the remote host
      (if (not remote-megatest)(set! remote-megatest local-megatest)) ;; "megatest"))
      (set! mt-bindir-path (pathname-directory remote-megatest))
      ;; (if launcher (set! launcher (string-split launcher)))           ;; yuk!
      ;; set up the run work area for this test
      (if (and (args:get-arg "-preclean") ;; user has requested to preclean for this run
	       (not (member (db:test-get-rundir testinfo)(list "n/a" "/tmp/badname")))) ;; n/a is a placeholder and thus not a read dir
	  (begin
	    (debug:print-info 0 *default-log-port* "attempting to preclean directory " (db:test-get-rundir testinfo) " for test " test-name "/" item-path)
	    (runs:remove-test-directory testinfo 'remove-data-only))) ;; remove data only, do not perturb the record
      
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
					(list 'mt-bindir-path mt-bindir-path))))))))
      
      ;; clean out step records from previous run if they exist
      ;; (rmt:delete-test-step-records run-id test-id)
      ;; if the dir does not exist we may have a itempath where individual variables are a path, launch anyway
      (if (common:file-exists? work-area)
	  (change-directory work-area)) ;; so that log files from the launch process don't clutter the test dir
      (cond
       ;; ((and launcher hosts) ;; must be using ssh hostname
       ;;    (set! fullcmd (append launcher (car hosts)(list remote-megatest "-m" test-sig "-execute" cmdparms) debug-param)))
       ;; (set! fullcmd (append launcher (car hosts)(list remote-megatest test-sig "-execute" cmdparms))))
       (launcher
	(set! fullcmd (append launcher (list remote-megatest "-m" test-sig "-execute" cmdparms) debug-param)))
       ;; (set! fullcmd (append launcher (list remote-megatest test-sig "-execute" cmdparms))))
       (else







|







1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
					(list 'mt-bindir-path mt-bindir-path))))))))
      
      ;; clean out step records from previous run if they exist
      ;; (rmt:delete-test-step-records run-id test-id)
      ;; if the dir does not exist we may have a itempath where individual variables are a path, launch anyway
      (if (common:file-exists? work-area)
	  (change-directory work-area)) ;; so that log files from the launch process don't clutter the test dir
      (cond       
       ;; ((and launcher hosts) ;; must be using ssh hostname
       ;;    (set! fullcmd (append launcher (car hosts)(list remote-megatest "-m" test-sig "-execute" cmdparms) debug-param)))
       ;; (set! fullcmd (append launcher (car hosts)(list remote-megatest test-sig "-execute" cmdparms))))
       (launcher
	(set! fullcmd (append launcher (list remote-megatest "-m" test-sig "-execute" cmdparms) debug-param)))
       ;; (set! fullcmd (append launcher (list remote-megatest test-sig "-execute" cmdparms))))
       (else
1618
1619
1620
1621
1622
1623
1624
















		     (read-symbolic-link (conc "/proc/" pid "/cwd"))
		     #f)))
    ;; now wait on that process if all is correct
    ;; periodically update the db with runtime
    ;; when the process exits look at the db, if still RUNNING after 10 seconds set
    ;; state/status appropriately
    (process-wait pid)))























>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
		     (read-symbolic-link (conc "/proc/" pid "/cwd"))
		     #f)))
    ;; now wait on that process if all is correct
    ;; periodically update the db with runtime
    ;; when the process exits look at the db, if still RUNNING after 10 seconds set
    ;; state/status appropriately
    (process-wait pid)))


 ;; (lock-key        (conc "test-" test-id))
	;; (got-lock        (let loop ((lock        (rmt:no-sync-get-lock lock-key))
	;; 			     (expire-time (+ (current-seconds) 15))) ;; give up on getting the lock and steal it after 15 seconds
	;; 		    (if (car lock)
	;; 			#t
	;; 			(if (> (current-seconds) expire-time)
	;; 			    (begin
	;; 			      (debug:print-info 0 *default-log-port* "Timed out waiting for a lock to launch test " keyvals " " runname " " test-name " " test-path)
	;; 			      (rmt:no-sync-del! lock-key) ;; destroy the lock
	;; 			      (loop (rmt:no-sync-get-lock lock-key) expire-time)) ;; 
	;; 			    (begin
	;; 			      (thread-sleep! 1)
	;; 			      (loop (rmt:no-sync-get-lock lock-key) expire-time))))))
	 

Modified rmt.scm from [39d97c528a] to [da397f2c5b].

925
926
927
928
929
930
931









932
933
934
935
936
937
938
  (rmt:send-receive 'no-sync-get/default #f `(,var ,default)))

(define (rmt:no-sync-del! var)
  (rmt:send-receive 'no-sync-del! #f `(,var)))

(define (rmt:no-sync-get-lock keyname)
  (rmt:send-receive 'no-sync-get-lock #f `(,keyname)))










;;======================================================================
;; A R C H I V E S
;;======================================================================

(define (rmt:archive-get-allocations  testname itempath dneeded)
  (rmt:send-receive 'archive-get-allocations #f (list testname itempath dneeded)))







>
>
>
>
>
>
>
>
>







925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
  (rmt:send-receive 'no-sync-get/default #f `(,var ,default)))

(define (rmt:no-sync-del! var)
  (rmt:send-receive 'no-sync-del! #f `(,var)))

(define (rmt:no-sync-get-lock keyname)
  (rmt:send-receive 'no-sync-get-lock #f `(,keyname)))

(define (rmt:no-sync-add-job host-type vars-list exekey)
  (rmt:send-receive 'no-sync-add-job #f `(,host-type ,vars-list ,exekey)))

(define (rmt:no-sync-take-job host-type)
  (rmt:send-receive 'no-sync-take-job #f `(,host-type)))

(define (rmt:no-sync-job-records-clean)
  (rmt:set-receive 'no-sync-job-records-clean #f '()))

;;======================================================================
;; A R C H I V E S
;;======================================================================

(define (rmt:archive-get-allocations  testname itempath dneeded)
  (rmt:send-receive 'archive-get-allocations #f (list testname itempath dneeded)))