Megatest

Check-in [bd65c1e661]
Login
Overview
Comment:wip
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | v1.7001-multi-db-wip2 | v1.7001-multi-db-rb01
Files: files | file ages | folders
SHA1: bd65c1e66194dce5a99a9ca31567c1961f4c640e
User & Date: matt on 2022-04-11 21:43:16
Other Links: branch diff | manifest | tags
Context
2022-04-12
07:15
no-sync db and db open locking working. Closed-Leaf check-in: a82e2bb971 user: matt tags: v1.7001-multi-db-wip2, v1.7001-multi-db-rb01
2022-04-11
21:43
wip check-in: bd65c1e661 user: matt tags: v1.7001-multi-db-wip2, v1.7001-multi-db-rb01
2022-04-07
07:04
wip check-in: 5209afd099 user: matt tags: v1.7001-multi-db-rb01
Changes

Modified api.scm from [fcdb3b5c5f] to [b65cdceb6b].

144
145
146
147
148
149
150

151
152
153
154
155
156
157

;; These are called by the server on recipt of /api calls
;;    - keep it simple, only return the actual result of the call, i.e. no meta info here
;;
;;    - returns #( flag result )
;;
(define (api:execute-requests dbstruct dat)

  (handle-exceptions
   exn
   (let ((call-chain (get-call-chain)))
     (debug:print 0 *default-log-port* "WARNING: api:execute-requests received an exception from peer, dat=" dat ", exn=" exn)
     (print-call-chain (current-error-port))
     (debug:print 0 *default-log-port* " message: "  ((condition-property-accessor 'exn 'message) exn))       
     (vector #f (vector exn call-chain dat))) ;; return some stuff for debug if an exception happens







>







144
145
146
147
148
149
150
151
152
153
154
155
156
157
158

;; These are called by the server on recipt of /api calls
;;    - keep it simple, only return the actual result of the call, i.e. no meta info here
;;
;;    - returns #( flag result )
;;
(define (api:execute-requests dbstruct dat)
  (db:open-no-sync-db) ;; sets *no-sync-db*
  (handle-exceptions
   exn
   (let ((call-chain (get-call-chain)))
     (debug:print 0 *default-log-port* "WARNING: api:execute-requests received an exception from peer, dat=" dat ", exn=" exn)
     (print-call-chain (current-error-port))
     (debug:print 0 *default-log-port* " message: "  ((condition-property-accessor 'exn 'message) exn))       
     (vector #f (vector exn call-chain dat))) ;; return some stuff for debug if an exception happens

Modified db.scm from [40daf428a9] to [5c6fcf9a46].

210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
	 (use-mutex (> *api-process-request-count* 25))) ;; was 25
    (if (and use-mutex
	     (common:low-noise-print 120 "over-50-parallel-api-requests"))
	(debug:print-info 0 *default-log-port* *api-process-request-count* " parallel api requests being processed in process " (current-process-id) ", throttling access"))
    (if (common:low-noise-print 600 (conc "parallel-api-requests" *max-api-process-requests*))
	(debug:print-info 2 *default-log-port* "Parallel api request count: " *api-process-request-count* " max parallel requests: " *max-api-process-requests*))
    (condition-case
     (begin
       (if use-mutex (mutex-lock! *db-with-db-mutex*))
       (let ((res (apply proc dbdat db params)))
	 (if use-mutex (mutex-unlock! *db-with-db-mutex*))
	 ;; (if (vector? dbstruct)(db:done-with dbstruct run-id r/w))
	 (if dbdat (stack-push! (dbr:subdb-dbstack subdb) dbdat))
	 res))
     (exn (io-error)
	  (db:generic-error-printout exn "ERROR: i/o error with " fname ". Check permissions, disk space etc. and try again."))
     (exn (corrupt)
	  (db:generic-error-printout exn "ERROR: database " fname " is corrupt. Repair it to proceed."))
     (exn (busy)
	  (db:generic-error-printout exn "ERROR: database " fname
				     " is locked. Try copying to another location, remove original and copy back."))
     (exn (permission)(db:generic-error-printout exn "ERROR: database " fname " has some permissions problem."))
     (exn ()
	  (db:generic-error-printout exn "ERROR: Unknown error with database " fname " message: "
		       ((condition-property-accessor 'exn 'message) exn))))))
      

;; NB// #f => return dbdir only
;;      (was planned to be;  zeroth db with name=main.db)
;; 
;; If run-id is #f return to create and retrieve the path where the db will live.
;;







|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|







210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
	 (use-mutex (> *api-process-request-count* 25))) ;; was 25
    (if (and use-mutex
	     (common:low-noise-print 120 "over-50-parallel-api-requests"))
	(debug:print-info 0 *default-log-port* *api-process-request-count* " parallel api requests being processed in process " (current-process-id) ", throttling access"))
    (if (common:low-noise-print 600 (conc "parallel-api-requests" *max-api-process-requests*))
	(debug:print-info 2 *default-log-port* "Parallel api request count: " *api-process-request-count* " max parallel requests: " *max-api-process-requests*))
    (condition-case
	(begin
	  (if use-mutex (mutex-lock! *db-with-db-mutex*))
	  (let ((res (apply proc dbdat db params)))
	    (if use-mutex (mutex-unlock! *db-with-db-mutex*))
	    ;; (if (vector? dbstruct)(db:done-with dbstruct run-id r/w))
	    (if dbdat (stack-push! (dbr:subdb-dbstack subdb) dbdat))
	    res))
      (exn (io-error)
	   (db:generic-error-printout exn "ERROR: i/o error with " fname ". Check permissions, disk space etc. and try again."))
      (exn (corrupt)
	   (db:generic-error-printout exn "ERROR: database " fname " is corrupt. Repair it to proceed."))
      (exn (busy)
	   (db:generic-error-printout exn "ERROR: database " fname
				      " is locked. Try copying to another location, remove original and copy back."))
      (exn (permission)(db:generic-error-printout exn "ERROR: database " fname " has some permissions problem."))
      (exn ()
	   (db:generic-error-printout exn "ERROR: Unknown error with database " fname " message: "
				      ((condition-property-accessor 'exn 'message) exn))))))
      

;; NB// #f => return dbdir only
;;      (was planned to be;  zeroth db with name=main.db)
;; 
;; If run-id is #f return to create and retrieve the path where the db will live.
;;
2103
2104
2105
2106
2107
2108
2109

















2110
2111
2112
2113
2114
2115
2116
  (db:with-db dbstruct #f #t 
	      (lambda (dbdat db)
		(sqlite3:execute db "DELETE FROM metadat WHERE var=?;" var))))

;;======================================================================
;; no-sync.db - small bits of data to be shared between servers
;;======================================================================


















(define (db:open-no-sync-db)
  (dbfile:open-no-sync-db (db:dbfile-path)))

(define (db:no-sync-close-db db stmt-cache)
  (db:safely-close-sqlite3-db db stmt-cache))








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







2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
  (db:with-db dbstruct #f #t 
	      (lambda (dbdat db)
		(sqlite3:execute db "DELETE FROM metadat WHERE var=?;" var))))

;;======================================================================
;; no-sync.db - small bits of data to be shared between servers
;;======================================================================

(define (db:no-sync-db db-in)
  (if db-in
      db-in
      (if *no-sync-db*
	  *no-sync-db*
	  (begin
	    (mutex-lock! *db-access-mutex*)
	    (let ((dbpath (common:get-db-tmp-area))
		  (db     (dbfile:open-no-sync-db dbpath)))
	      (set! *no-sync-db* db)
	      (mutex-unlock! *db-access-mutex*)
	      db)))))

(define (with-no-sync-db proc)
  (let* ((db  (db:no-sync-db *no-sync-db*)))
    (proc db)))

(define (db:open-no-sync-db)
  (dbfile:open-no-sync-db (db:dbfile-path)))

(define (db:no-sync-close-db db stmt-cache)
  (db:safely-close-sqlite3-db db stmt-cache))

5007
5008
5009
5010
5011
5012
5013

5014




5015



5016
5017
5018
5019
5020
5021
5022
	 (dbfiles        (glob (conc tmp-area"/.db/*.db")))
	 (sync-durations (make-hash-table)))
    ;; (debug:print-info 0 *default-log-port* "lock-and-sync, dbfiles: "dbfiles)
    (for-each
     (lambda (file)
       (let* ((fname (conc (pathname-file file) ".db"))
	      (fulln (conc *toppath*"/.db/"fname))

	      (time1 (file-modification-time file))




	      (time2 (file-modification-time fulln))



	      (changed (> time1 time2))
	      (do-cp (cond
		      ((not (file-exists? fulln)) ;; shouldn't happen, but this might recover
		       (debug:print-info 0 *default-log-port* "File "fulln" not found! Copying "fname" to "fulln)
		       #t)
		      (changed ;; (and changed
		       ;; (> (- (current-seconds) time1) 3)) ;; if file is changed and three seconds have passed.







>
|
>
>
>
>
|
>
>
>







5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
5043
5044
5045
5046
5047
	 (dbfiles        (glob (conc tmp-area"/.db/*.db")))
	 (sync-durations (make-hash-table)))
    ;; (debug:print-info 0 *default-log-port* "lock-and-sync, dbfiles: "dbfiles)
    (for-each
     (lambda (file)
       (let* ((fname (conc (pathname-file file) ".db"))
	      (fulln (conc *toppath*"/.db/"fname))
	      (time1 (if (file-exists? file)
			 (file-modification-time file)
			 (begin
			   (debug:print-info 0 *default-log-port* "Sync - I do not see file "file)
			   1)))
	      (time2 (if (file-exists? fulln)
			 (file-modification-time fulln)
			 (begin
			   (debug:print-info 0 *default-log-port* "Sync - I do not see file "fulln)
			   0)))
	      (changed (> time1 time2))
	      (do-cp (cond
		      ((not (file-exists? fulln)) ;; shouldn't happen, but this might recover
		       (debug:print-info 0 *default-log-port* "File "fulln" not found! Copying "fname" to "fulln)
		       #t)
		      (changed ;; (and changed
		       ;; (> (- (current-seconds) time1) 3)) ;; if file is changed and three seconds have passed.

Modified dbfile.scm from [c5ec4d8a51] to [13bb3140b5].

442
443
444
445
446
447
448
449



450

451

452

453



454
455
456

457




458
459
460

461

462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
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
;; no-sync.db - small bits of data to be shared between servers
;;======================================================================

;; 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)



  (if db-in

      db-in

      (if *no-sync-db*

	  *no-sync-db*



	  (begin
	    (mutex-lock! *db-access-mutex*)
	    (let ((db (dbfile:open-no-sync-db)))

	      (set! *no-sync-db* db)




	      (mutex-unlock! *db-access-mutex*)
	      db)))))


(define (dbfile:open-no-sync-db dbpath)

  (let* (;; (dbpath (db:dbfile-path))
	 (dbname (conc dbpath "/no-sync.db"))
	 (db-exists (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))

(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
        (let ((newres (if (string? res)
			  (string->number res)
			  #f)))
          (if newres
              newres
              res))
        res)))

;; transaction protected lock aquisition
;; either:
;;    fails    returns  (#f . lock-creation-time)
;;    succeeds (returns (#t . lock-creation-time)
;; use (db:no-sync-del! db keyname) to release the lock
;;
(define (db:no-sync-get-lock db-in keyname)
  (let ((db (db:no-sync-db db-in)))
    (sqlite3:with-transaction
     db
     (lambda ()
       (handle-exceptions
	   exn
	 (let ((lock-time (current-seconds)))
	   ;; (debug:print-info 2 *default-log-port* "db:no-sync-get-lock keyname=" keyname ", lock-time=" lock-time ", exn=" exn)
	   (sqlite3:execute db "INSERT INTO no_sync_metadat (var,val) VALUES(?,?);" keyname lock-time)
	   `(#t . ,lock-time))
	 `(#f . ,(sqlite3:first-result db "SELECT val FROM no_sync_metadat WHERE var=?;" keyname)))))))


;;======================================================================
;; file utils
;;======================================================================

;;======================================================================







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

>

>
|
|

|





|



|


|






|

















|
<
|
|
|
|
|




|







442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463

464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
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
524
525
526
527
528
529
530
531
532
533
534
535
;; no-sync.db - small bits of data to be shared between servers
;;======================================================================

;; 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 (dbfile:cautious-open-database fname #!optional (tries-left 5))
  (let* ((retry (lambda ()
		  (thread-sleep! 0.5)
		  (if (> tries-left 0)
		      (dbfile:cautious-open-database fname (- tries-left 1))))))
    (condition-case
	(sqlite3:open-database fname)
      (exn (io-error)
	   (dbfile:print-err exn "ERROR: i/o error with " fname ". Check permissions, disk space etc. and try again.")
	   (retry))
      (exn (corrupt)
	   (dbfile:print-err exn "ERROR: database " fname " is corrupt. Repair it to proceed.")
	   (retry))
      (exn (busy)

	   (dbfile:print-err exn "ERROR: database " fname
			     " is locked. Try copying to another location, remove original and copy back.")
	   (retry))
      (exn (permission)(dbfile:print-err exn "ERROR: database " fname " has some permissions problem.")
	   (retry))
      (exn ()
	   (dbfile:print-err exn "ERROR: Unknown error with database " fname " message: "
			     ((condition-property-accessor 'exn 'message) exn))
	   (retry)))))


(define (dbfile:open-no-sync-db dbpath)
  (if (not (file-exists? dbpath))
      (create-directory dbpath #t))
  (let* ((dbname    (conc dbpath "/no-sync.db"))
	 (db-exists (file-exists? dbname))
	 (db        (dbfile:cautious-open-database dbname))) ;; (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))

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

(define (db:no-sync-del! db var)
  (sqlite3:execute 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
     "SELECT val FROM no_sync_metadat WHERE var=?;"
     var)
    (if res
        (let ((newres (if (string? res)
			  (string->number res)
			  #f)))
          (if newres
              newres
              res))
        res)))

;; transaction protected lock aquisition
;; either:
;;    fails    returns  (#f . lock-creation-time)
;;    succeeds (returns (#t . lock-creation-time)
;; use (db:no-sync-del! db keyname) to release the lock
;;
(define (db:no-sync-get-lock db keyname)

  (sqlite3:with-transaction
   db
   (lambda ()
     (handle-exceptions
	 exn
	 (let ((lock-time (current-seconds)))
	   ;; (debug:print-info 2 *default-log-port* "db:no-sync-get-lock keyname=" keyname ", lock-time=" lock-time ", exn=" exn)
	   (sqlite3:execute db "INSERT INTO no_sync_metadat (var,val) VALUES(?,?);" keyname lock-time)
	   `(#t . ,lock-time))
       `(#f . ,(sqlite3:first-result db "SELECT val FROM no_sync_metadat WHERE var=?;" keyname))))))


;;======================================================================
;; file utils
;;======================================================================

;;======================================================================

Modified http-transport.scm from [df8bc9e37e] to [4a4c885dbb].

401
402
403
404
405
406
407

408
409
410
411
412
413
414
;;
(define (http-transport:keep-running) 
  ;; if none running or if > 20 seconds since 
  ;; server last used then start shutdown
  ;; This thread waits for the server to come alive
  (debug:print-info 0 *default-log-port* "Starting the sync-back, keep alive thread in server")
  (let* ((sdat              #f)

	 (tmp-area          (common:get-db-tmp-area))
	 (started-file      (conc tmp-area "/.server-started"))
	 (server-start-time (current-seconds))
	 (server-info (let loop ((start-time (current-seconds))
				 (changed    #t)
				 (last-sdat  "not this"))
                        (begin ;; let ((sdat #f))







>







401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
;;
(define (http-transport:keep-running) 
  ;; if none running or if > 20 seconds since 
  ;; server last used then start shutdown
  ;; This thread waits for the server to come alive
  (debug:print-info 0 *default-log-port* "Starting the sync-back, keep alive thread in server")
  (let* ((sdat              #f)
	 (no-sync-db        (db:open-no-sync-db))
	 (tmp-area          (common:get-db-tmp-area))
	 (started-file      (conc tmp-area "/.server-started"))
	 (server-start-time (current-seconds))
	 (server-info (let loop ((start-time (current-seconds))
				 (changed    #t)
				 (last-sdat  "not this"))
                        (begin ;; let ((sdat #f))

Modified tasks.scm from [19e9ab848e] to [e54ee89bd7].

512
513
514
515
516
517
518
519
520
521
522
523


524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
      exn
      '()
      (sqlite3:first-row db "SELECT id,action,owner,state,target,name,testpatt,keylock,params WHERE
                               params LIKE ? AND state LIKE ? AND action LIKE ? AND testpatt LIKE ?;"
			 param-key state-patt action-patt test-patt)))))

(define (tasks:find-task-queue-records dbstruct target run-name test-patt state-patt action-patt)
  ;; (handle-exceptions
  ;;  exn
  ;;  '()
  ;;  (sqlite3:first-row
  (let ((db (db:delay-if-busy (db:get-db dbstruct)))


	(res '()))
    (sqlite3:for-each-row 
     (lambda (a . b)
       (set! res (cons (cons a b) res)))
     db "SELECT id,action,owner,state,target,name,testpatt,keylock,params FROM tasks_queue 
           WHERE
              target = ? AND name = ? AND state LIKE ? AND action LIKE ? AND testpatt LIKE ?;"
     target run-name state-patt action-patt test-patt)
    res)) ;; )

;; kill any runner processes (i.e. processes handling -runtests) that match target/runname
;; 
;; do a remote call to get the task queue info but do the killing as self here.
;;
(define (tasks:kill-runner target run-name testpatt)
  (let ((records    (rmt:tasks-find-task-queue-records target run-name testpatt "running" "run-tests"))







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


|
|







512
513
514
515
516
517
518



519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
      exn
      '()
      (sqlite3:first-row db "SELECT id,action,owner,state,target,name,testpatt,keylock,params WHERE
                               params LIKE ? AND state LIKE ? AND action LIKE ? AND testpatt LIKE ?;"
			 param-key state-patt action-patt test-patt)))))

(define (tasks:find-task-queue-records dbstruct target run-name test-patt state-patt action-patt)



  (db:with-db
   dbstruct
   #f #f
   (lambda (dbdat db)
     (let ((res '()))
       (sqlite3:for-each-row 
	(lambda (a . b)
	  (set! res (cons (cons a b) res)))
	db "SELECT id,action,owner,state,target,name,testpatt,keylock,params FROM tasks_queue 
           WHERE
              target = ? AND name = ? AND state LIKE ? AND action LIKE ? AND testpatt LIKE ?;"
	target run-name state-patt action-patt test-patt)
       res))))

;; kill any runner processes (i.e. processes handling -runtests) that match target/runname
;; 
;; do a remote call to get the task queue info but do the killing as self here.
;;
(define (tasks:kill-runner target run-name testpatt)
  (let ((records    (rmt:tasks-find-task-queue-records target run-name testpatt "running" "run-tests"))

Modified tests/simplerun/Makefile from [115e15e0c2] to [d958b22b7d].

1
2
3
4
5

cleanup :
	killall mtest dboard -v -9 || true
	rm -rf *.log *.bak NB* logs/* .meta .db ../simpleruns/* lt




|

1
2
3
4
5

cleanup :
	killall mtest dboard -v -9 || true
	rm -rf *.log *.bak NB* logs/* .meta .db /tmp/$(USER)/megatest_localdb/simplerun ../simpleruns/* lt