Megatest

Diff
Login

Differences From Artifact [ac12c8febe]:

To Artifact [cb9f17b39f]:


56
57
58
59
60
61
62
63

64
65
66
67
68
69
70
56
57
58
59
60
61
62

63
64
65
66
67
68
69
70







-
+







	   (set! res adr)))
     (vector->list (hostinfo-addresses (hostname->hostinfo hostname))))
    (string-intersperse 
     (map number->string
	  (u8vector->list
	   (if res res (hostname->ip hostname)))) ".")))

(define (http-transport:run hostn)
(define (http-transport:run hostn run-id server-id)
  (debug:print 2 "Attempting to start the server ...")
  (if (not *toppath*)
      (if (not (setup-for-run))
	  (begin
	    (debug:print 0 "ERROR: cannot find megatest.config, cannot start server, exiting")
	    (exit))))
  (let* (;; (iface           (if (string=? "-" hostn)
139
140
141
142
143
144
145
146

147
148
149
150

151
152
153
154
155
156
157
158
159
160
161

162
163
164

165
166

167
168
169


170
171
172
173
174
175
176
139
140
141
142
143
144
145

146
147
148
149

150
151
152
153
154
155
156
157
158



159
160
161

162


163
164


165
166
167
168
169
170
171
172
173







-
+



-
+








-
-
-
+


-
+
-
-
+

-
-
+
+







				   (send-response body: "hey there!\n"
						  headers: '((content-type text/plain))))
				  ((equal? (uri-path (request-uri (current-request))) 
					   '(/ "hey"))
				   (send-response body: "hey there!\n"
						  headers: '((content-type text/plain))))
				  (else (continue))))))))
    (http-transport:try-start-server ipaddrstr start-port)))
    (http-transport:try-start-server ipaddrstr start-port server-id)))

;; This is recursively run by http-transport:run until sucessful
;;
(define (http-transport:try-start-server ipaddrstr portnum)
(define (http-transport:try-start-server ipaddrstr portnum server-id)
  (handle-exceptions
   exn
   (begin
     (print-error-message exn)
     (if (< portnum 9000)
	 (begin 
	   (debug:print 0 "WARNING: failed to start on portnum: " portnum ", trying next port")
	   (thread-sleep! 0.1)
	   ;; (open-run-close tasks:remove-server-records tasks:open-db)
	   (open-run-close tasks:server-delete tasks:open-db ipaddrstr portnum)
	   (http-transport:try-start-server ipaddrstr (+ portnum 1)))
	   (http-transport:try-start-server ipaddrstr (+ portnum 1) server-id))
	 (print "ERROR: Tried and tried but could not start the server")))
   ;; any error in following steps will result in a retry
   (set! *runremote* (list ipaddrstr portnum))
   (set! *server-info* (list ipaddrstr portnum))
   ;; (open-run-close tasks:remove-server-records tasks:open-db)
   (open-run-close tasks:server-register 
   (open-run-close tasks:server-set-interface-port 
		   tasks:open-db 
		   (current-process-id)
		   ipaddrstr portnum 0 'startup 'http)
		   server-id 
		   ipaddrstr portnum)
   (debug:print 1 "INFO: Trying to start server on " ipaddrstr ":" portnum)
   ;; This starts the spiffy server
   ;; NEED WAY TO SET IP TO #f TO BIND ALL
   (start-server bind-address: ipaddrstr port: portnum)
   (open-run-close tasks:server-delete tasks:open-db ipaddrstr portnum)
   (debug:print 1 "INFO: server has been stopped")))

294
295
296
297
298
299
300
301

302
303
304
305
306
307
308
309
310
311
312
313
314


315

316
317
318
319
320
321
322
291
292
293
294
295
296
297

298
299
300
301
302
303
304
305
306
307
308
309

310
311
312

313
314
315
316
317
318
319
320







-
+











-

+
+
-
+







	   (debug:print-info 11 "match=" match)
	   (let ((final (cadr match)))
	     (debug:print-info 11 "final=" final)
	     final)))))))

;; Send "cmd" with json payload "params" to serverdat and receive result
;;
(define (http-transport:client-api-send-receive serverdat cmd params #!key (numretries 30))
(define (http-transport:client-api-send-receive run-id serverdat cmd params #!key (numretries 30))
  (let* ((fullurl    (if (list? serverdat)
			 (cadddr serverdat) ;; this is the uri for /api
			 (begin
			   (debug:print 0 "FATAL ERROR: http-transport:client-send-receive called with no server info")
			   (exit 1))))
	 (res        #f))
    (handle-exceptions
     exn
     (begin
       ;; TODO: Send this output to a log file so it isn't lost when running as daemon
       (print "ERROR IN http-transport:client-send-receive " ((condition-property-accessor 'exn 'message) exn))
       (thread-sleep! 2)
       (if (> numretries 0)
	   (begin
	     (if (> (random 100) 80)(server:ensure-running run-id)) ;; every so often try starting a server
	   (http-transport:client-api-send-receive serverdat cmd params numretries: (- numretries 1))))
	     (http-transport:client-api-send-receive run-id serverdat cmd params numretries: (- numretries 1)))))
     (begin
       (debug:print-info 11 "fullurl=" fullurl "\n")
       ;; set up the http-client here
       (max-retry-attempts 5)
       ;; consider all requests indempotent
       (retry-request? (lambda (request)
			 #t))   ;;  		 (thread-sleep! (/ (if (> numretries 100) 100 numretries) 10))
371
372
373
374
375
376
377
378

379
380
381
382
383
384


385
386
387
388
389

390
391
392
393
394
395
396
397
398
399
400
401
402

403
404
405
406



407
408
409

410
411



412
413
414


415

416
417
418
419
420
421
422
423
424
425
426
427
428

429
430
431
432
433
434
435
369
370
371
372
373
374
375

376
377
378
379
380


381
382
383
384
385
386

387
388
389
390
391




392
393
394
395

396
397
398
399

400
401
402
403
404

405
406

407
408
409
410
411
412
413
414

415
416
417
418
419


420
421
422
423
424
425

426
427
428
429
430
431
432
433







-
+




-
-
+
+




-
+




-
-
-
-




-
+



-
+
+
+


-
+

-
+
+
+



+
+
-
+




-
-






-
+







	 (thread-start! th1)
	 (thread-start! th2)
	 (thread-join! th1)
	 (thread-terminate! th2)
	 (debug:print-info 11 "got res=" res)
	 res)))))

(define (http-transport:client-connect iface port)
(define (http-transport:client-connect run-id iface port)
  (let* ((login-res   #f)
	 (uri-dat     (make-request method: 'POST uri: (uri-reference (conc "http://" iface ":" port "/ctrl"))))
	 (uri-api-dat (make-request method: 'POST uri: (uri-reference (conc "http://" iface ":" port "/api"))))
	 (serverdat   (list iface port uri-dat uri-api-dat)))
    (set! *runremote* serverdat) ;; may or may not be good ...
    (set! login-res (rmt:login))
    (hash-table-set! *runremote* run-id serverdat) ;; may or may not be good ...
    (set! login-res (rmt:login run-id))
    (if (and (list? login-res)
	     (car login-res))
	(begin
	  (debug:print-info 2 "Logged in and connected to " iface ":" port)
	  (set! *runremote* serverdat)
	  (hash-table-set! *runremote* run-id serverdat)
	  serverdat)
	(begin
	  (debug:print-info 0 "ERROR: Failed to login or connect to " iface ":" port)
	  (exit 1)))))
;; 	  (set! *runremote* #f)
;; 	  (set! *transport-type* 'fs)
;; 	  #f))))


;; run http-transport:keep-running in a parallel thread to monitor that the db is being 
;; used and to shutdown after sometime if it is not.
;;
(define (http-transport:keep-running)
(define (http-transport:keep-running server-id)
  ;; if none running or if > 20 seconds since 
  ;; server last used then start shutdown
  ;; This thread waits for the server to come alive
  (let* ((server-info (let loop ()
  (let* ((server-info (let loop ((start-time (current-seconds))
				 (changed    #t)
				 (last-sdat  "not this"))
                        (let ((sdat #f))
                          (mutex-lock! *heartbeat-mutex*)
                          (set! sdat *runremote*)
                          (set! sdat *server-info*)
                          (mutex-unlock! *heartbeat-mutex*)
                          (if sdat
                          (if (and sdat
				   (not changed)
				   (> (- (current-seconds) start-time) 2))
			      sdat
                              (begin
                                (sleep 4)
                                (loop start-time
				      (equal? sdat last-sdat)
                                (loop))))))
				      sdat))))))
         (iface       (car server-info))
         (port        (cadr server-info))
         (last-access 0)
	 (tdb         (tasks:open-db))
	 (spid        ;;(open-run-close tasks:server-get-server-id tasks:open-db #f iface port #f))
	   (tasks:server-get-server-id tdb #f iface port #f))
	 (server-timeout (let ((tmo (config-lookup  *configdat* "server" "timeout")))
			   (if (and (string? tmo)
				    (string->number tmo))
			       (* 60 60 (string->number tmo))
			       ;; default to three days
			       (* 3 24 60 60)))))
    (debug:print-info 2 "server-timeout: " server-timeout ", server pid: " spid " on " iface ":" port)
    (tasks:server-set-state! tdb server-id "running")
    (let loop ((count 0))
      ;; Use this opportunity to sync the inmemdb to db
      (let ((start-time (current-milliseconds))
	    (sync-time  #f)
	    (rem-time   #f))
	(if *inmemdb* (db:sync-touched *inmemdb* force-sync: #t))
	(set! sync-time  (- (current-milliseconds) start-time))
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
441
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
536
537







-
+



-
+



-
+
-



-
+




















-
-
+
+
















+



-
+
+








-
+
-
-
-
-
+
+
+
+





+
+
-
-
+
+
+
+

-
+







      ;; (thread-sleep! 4) ;; no need to do this very often

      (if (< count 1) ;; 3x3 = 9 secs aprox
	  (loop (+ count 1)))
      
      ;; Check that iface and port have not changed (can happen if server port collides)
      (mutex-lock! *heartbeat-mutex*)
      (set! sdat *runremote*)
      (set! sdat *server-info*)
      (mutex-unlock! *heartbeat-mutex*)
      
      (if (or (not (equal? sdat (list iface port)))
	      (not spid))
	      (not server-id))
	  (begin 
	    (debug:print-info 0 "interface changed, refreshing iface and port info")
	    (set! iface (car sdat))
	    (set! port  (cadr sdat))
	    (set! port  (cadr sdat))))
	    (set! spid  (tasks:server-get-server-id tdb #f iface port #f))))
      
      ;; NOTE: Get rid of this mechanism! It really is not needed...
      ;; (open-run-close tasks:server-update-heartbeat tasks:open-db spid)
      (tasks:server-update-heartbeat tdb spid)
      (tasks:server-update-heartbeat tdb server-id)
      
      ;; (if ;; (or (> numrunning 0) ;; stay alive for two days after last access

      ;; Transfer *last-db-access* to last-access to use in checking that we are still alive
      (mutex-lock! *heartbeat-mutex*)
      (set! last-access *last-db-access*)
      (mutex-unlock! *heartbeat-mutex*)

      ;; (debug:print 11 "last-access=" last-access ", server-timeout=" server-timeout)
      (if (and *server-run*
	       (> (+ last-access server-timeout)
		  (current-seconds)))
	  (begin
	    (debug:print-info 0 "Server continuing, seconds since last db access: " (- (current-seconds) last-access))
	    (loop 0))
	  (begin
	    (debug:print-info 0 "Starting to shutdown the server.")
	    ;; need to delete only *my* server entry (future use)
	    (set! *time-to-exit* #t)
	    (if *inmemdb* (db:sync-touched *inmemdb* force-sync: #t))
	    (open-run-close tasks:server-deregister-self tasks:open-db (get-host-name))
	    (thread-sleep! 1)
	    ( tasks:server-set-state! tdb server-id "shutting-down")
	    (thread-sleep! 5)
	    (debug:print-info 0 "Max cached queries was    " *max-cache-size*)
	    (debug:print-info 0 "Number of cached writes   " *number-of-writes*)
	    (debug:print-info 0 "Average cached write time "
			      (if (eq? *number-of-writes* 0)
				  "n/a (no writes)"
				  (/ *writes-total-delay*
				     *number-of-writes*))
			      " ms")
	    (debug:print-info 0 "Number non-cached queries "  *number-non-write-queries*)
	    (debug:print-info 0 "Average non-cached time   "
			      (if (eq? *number-non-write-queries* 0)
				  "n/a (no queries)"
				  (/ *total-non-write-delay* 
				     *number-non-write-queries*))
			      " ms")
	    (debug:print-info 0 "Server shutdown complete. Exiting")
	    (tasks:server-delete-record! tdb server-id)
	    (exit))))))

;; all routes though here end in exit ...
(define (http-transport:launch)
(define (http-transport:launch run-id)
  (set! *run-id*   run-id)
  (if (not *toppath*)
      (if (not (setup-for-run))
	  (begin
	    (debug:print 0 "ERROR: cannot find megatest.config, exiting")
	    (exit))))
  (debug:print-info 2 "Starting the standalone server")
  (if (args:get-arg "-daemonize")
      (daemon:ize))
  (let ((hostinfo (open-run-close tasks:get-best-server tasks:open-db)))
  (let ((server-id (open-run-close tasks:server-lock-slot tasks:open-db run-id)))
    (debug:print 11 "http-transport:launch hostinfo=" hostinfo)
    ;; #(1 "143.182.207.24" 5736 -1 "http" 22771 "hostname")
    (if hostinfo
	(debug:print-info 2 "NOT starting new server, one is already running on " (vector-ref hostinfo 1) ":" (vector-ref hostinfo 2))
    (if (not server-id)
	(begin
	  (debug:print-info 2 "INFO: server pid=" (current-process-id) ", hostname=" (get-host-name) " not starting due to other candidates ahead in start queue")
	  (open-run-close tasks:server-delete-records-for-this-pid tasks:open-db))
	(if *toppath* 
	    (let* ((th2 (make-thread (lambda ()
				       (http-transport:run 
					(if (args:get-arg "-server")
					    (args:get-arg "-server")
					    "-")
					run-id
					    "-"))) "Server run"))
		   (th3 (make-thread http-transport:keep-running "Keep running")))
					server-id)) "Server run"))
		   (th3 (make-thread (lambda ()
				       (http-transport:keep-running server-id))
				     "Keep running")))
	      ;; Database connection
	      (set! *inmemdb*  (db:setup))
	      (set! *inmemdb*  (db:setup run-id))
	      (thread-start! th2)
	      (thread-start! th3)
	      (set! *didsomething* #t)
	      (thread-join! th2))
	    (debug:print 0 "ERROR: Failed to setup for megatest")))
    ;; (sdb:qry 'finalize)
    (exit)))