Megatest

Check-in [c5f14c055a]
Login
Overview
Comment:short and fast server start implemented (just needs extensive testing)
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | v1.80-revolution
Files: files | file ages | folders
SHA1: c5f14c055ab25cc86c9e808d78c2cd3928521843
User & Date: matt on 2023-11-27 18:40:41
Other Links: branch diff | manifest | tags
Context
2023-11-27
18:56
Bump version check-in: bca03ac0a6 user: matt tags: v1.80-revolution
18:40
short and fast server start implemented (just needs extensive testing) check-in: c5f14c055a user: matt tags: v1.80-revolution
17:06
new start up process implemented (but not tested) for main.db check-in: af90c933ed user: mrwellan tags: v1.80-revolution
Changes

Modified tcp-transportmod.scm from [1701bddb00] to [2aca24dc6e].

149
150
151
152
153
154
155
156



157
158
159
160
161
162
163
				   testsuite ;; (dbfile:testsuite-name)
				   (common:find-local-megatest)
				   run-id)))))
    (if conn
	(begin 
          (debug:print-info 2 *default-log-port* "already connected to a server")
           conn) ;; we are already connected to the server
	(let* ((sdat (tt:get-current-server-info ttdat dbfname)))



	  (match sdat
	    ((host port start-time server-id pid dbfname2 servinffile)
	     (assert (equal? dbfname dbfname2) "FATAL: read server info from wrong file.")
             (debug:print-info 2 *default-log-port* "no conn - in match servinffile:" servinffile)
	     (let* ((host-port (conc host":"port))
		    (conn (make-tt-conn
			   host: host







|
>
>
>







149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
				   testsuite ;; (dbfile:testsuite-name)
				   (common:find-local-megatest)
				   run-id)))))
    (if conn
	(begin 
          (debug:print-info 2 *default-log-port* "already connected to a server")
           conn) ;; we are already connected to the server
	(let* ((sdats (tt:get-server-info-sorted ttdat dbfname))
	       (sdat  (if (null? sdats)
			  #f
			  (car sdats))))
	  (match sdat
	    ((host port start-time server-id pid dbfname2 servinffile)
	     (assert (equal? dbfname dbfname2) "FATAL: read server info from wrong file.")
             (debug:print-info 2 *default-log-port* "no conn - in match servinffile:" servinffile)
	     (let* ((host-port (conc host":"port))
		    (conn (make-tt-conn
			   host: host
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
		    (let* ((curr-secs (current-seconds)))
		      ;; rm the (last server) would go here
		      (if (> (- curr-secs (tt-last-serv-start ttdat)) 10)
			  (begin
			    (tt-last-serv-start-set! ttdat curr-secs)
			    (server-start-proc))) ;; start server if 10 sec since last attempt
		      (thread-sleep! 1)
                      (debug:print-info 2 *default-log-port* "server ping result was neither running nor starting. Retrying connect")
		      (tt:client-connect-to-server ttdat dbfname run-id testsuite server-start-proc)))))))
	    (else ;; no good server found, if haven't started server in > 5 secs, start another
	     (if (> (- (current-seconds) (tt-last-serv-start ttdat)) 3) ;; BUG - grow this number really do not want to swamp the machine with servers
		 (begin
		   (debug:print-info 0 *default-log-port* "Starting server for "dbfname)
		   (server-start-proc)
		   (tt-last-serv-start-set! ttdat (current-seconds))







|







189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
		    (let* ((curr-secs (current-seconds)))
		      ;; rm the (last server) would go here
		      (if (> (- curr-secs (tt-last-serv-start ttdat)) 10)
			  (begin
			    (tt-last-serv-start-set! ttdat curr-secs)
			    (server-start-proc))) ;; start server if 10 sec since last attempt
		      (thread-sleep! 1)
                      (debug:print-info 0 *default-log-port* "server ping result was neither running nor starting. Retrying connect")
		      (tt:client-connect-to-server ttdat dbfname run-id testsuite server-start-proc)))))))
	    (else ;; no good server found, if haven't started server in > 5 secs, start another
	     (if (> (- (current-seconds) (tt-last-serv-start ttdat)) 3) ;; BUG - grow this number really do not want to swamp the machine with servers
		 (begin
		   (debug:print-info 0 *default-log-port* "Starting server for "dbfname)
		   (server-start-proc)
		   (tt-last-serv-start-set! ttdat (current-seconds))
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
		   ;; (debug:print 0 *default-log-port* "INFO: got corrupt data from server "host":"port", "res", for "dbfname", will try again.")
		   ;; (tt:handler ttdat cmd run-id params (+ attemptnum 1) readonly-mode dbfname testsuite mtexe)
		   )))))
	(begin
	  (thread-sleep! 1) ;; no conn yet set up, give it a rest and try again
	  (tt:handler ttdat cmd run-id params attemptnum readonly-mode dbfname testsuite mtexe server-start-proc)))))

(define (tt:bid-for-servership run-id)
  #f)

;; gets server info and appends path to server file
;; sorts by age, oldest first
;;
;; returns list of (host port startseconds server-id servinfofile)
;;
(define (tt:get-server-info-sorted ttdat dbfname)
  (let* ((areapath (tt-areapath ttdat))







<
<
<







314
315
316
317
318
319
320



321
322
323
324
325
326
327
		   ;; (debug:print 0 *default-log-port* "INFO: got corrupt data from server "host":"port", "res", for "dbfname", will try again.")
		   ;; (tt:handler ttdat cmd run-id params (+ attemptnum 1) readonly-mode dbfname testsuite mtexe)
		   )))))
	(begin
	  (thread-sleep! 1) ;; no conn yet set up, give it a rest and try again
	  (tt:handler ttdat cmd run-id params attemptnum readonly-mode dbfname testsuite mtexe server-start-proc)))))




;; gets server info and appends path to server file
;; sorts by age, oldest first
;;
;; returns list of (host port startseconds server-id servinfofile)
;;
(define (tt:get-server-info-sorted ttdat dbfname)
  (let* ((areapath (tt-areapath ttdat))
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
       (if (or (> (length sorted) 1)
	       (common:low-noise-print 120 "server info sorted"))
	   (debug:print 2 *default-log-port* "SERVER #"count": "(string-intersperse (map conc sorted) ", ")))
       (set! count (+ count 1)))
     sorted)
    sorted))
    
(define (tt:get-current-server-info ttdat dbfname)
  (assert (tt-areapath ttdat) "FATAL: areapath not set in ttdat.")
  ;;
  ;; TODO - replace most of below with tt;get-server-info-sorted
  ;;
  (let* ((areapath (tt-areapath ttdat))
	 (sfiles   (tt:find-server areapath dbfname))
	 (sdats    (filter car (map tt:server-get-info sfiles))) ;; first element is #f if the file disappeared while being read
	 (sorted   (sort sdats (lambda (a b)
				 (< (list-ref a 2)(list-ref b 2))))))
    (if (null? sorted)
	#f  ;; we'll want to wait until extra servers have exited
	(car sorted))))

(define (tt:send-receive ttdat conn cmd run-id params)
  (let* ((host-port (tt-conn-host-port conn)) ;; (conc (tt-conn-host conn)":"(tt-conn-port conn)))
	 (host      (tt-conn-host conn))
	 (port      (tt-conn-port conn))
	 (dat       (list cmd run-id params #f))) ;; no meta data yet
    (tt:send-receive-direct host port dat)))








<
<
<
<
<
<
<
<
<
<
<
<
<
<







339
340
341
342
343
344
345














346
347
348
349
350
351
352
       (if (or (> (length sorted) 1)
	       (common:low-noise-print 120 "server info sorted"))
	   (debug:print 2 *default-log-port* "SERVER #"count": "(string-intersperse (map conc sorted) ", ")))
       (set! count (+ count 1)))
     sorted)
    sorted))
    














(define (tt:send-receive ttdat conn cmd run-id params)
  (let* ((host-port (tt-conn-host-port conn)) ;; (conc (tt-conn-host conn)":"(tt-conn-port conn)))
	 (host      (tt-conn-host conn))
	 (port      (tt-conn-port conn))
	 (dat       (list cmd run-id params #f))) ;; no meta data yet
    (tt:send-receive-direct host port dat)))

479
480
481
482
483
484
485

486
487
488
489
490
491
492
;;
;; Server viability is checked in keep-running. Blindly start and run here.
;;
(define (tt:start-server areapath run-id dbfname-in handler keys)
  (assert areapath "FATAL: areapath not provided for tt:start-server")
  (let* ((ttdat   (make-tt areapath: areapath))
	 (dbfname (or dbfname-in (dbmod:run-id->dbfname run-id))))

    (let* ((dbstruct   (dbmod:open-dbmoddb areapath run-id dbfname (dbfile:db-init-proc) keys)))
      (tt-handler-set! ttdat (handler dbstruct))
      (let* ((servinf-created #f)
	     (tcp-thread      (make-thread
			       (lambda ()
				 ;; NOTE: tt-port and tt-host are set in connect-listener which is called under tt:start-tcp-server
				 (tt:start-tcp-server ttdat)) ;; start the tcp-server which applies handler to incoming data







>







465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
;;
;; Server viability is checked in keep-running. Blindly start and run here.
;;
(define (tt:start-server areapath run-id dbfname-in handler keys)
  (assert areapath "FATAL: areapath not provided for tt:start-server")
  (let* ((ttdat   (make-tt areapath: areapath))
	 (dbfname (or dbfname-in (dbmod:run-id->dbfname run-id))))
    (set! *server-info* ttdat)
    (let* ((dbstruct   (dbmod:open-dbmoddb areapath run-id dbfname (dbfile:db-init-proc) keys)))
      (tt-handler-set! ttdat (handler dbstruct))
      (let* ((servinf-created #f)
	     (tcp-thread      (make-thread
			       (lambda ()
				 ;; NOTE: tt-port and tt-host are set in connect-listener which is called under tt:start-tcp-server
				 (tt:start-tcp-server ttdat)) ;; start the tcp-server which applies handler to incoming data
568
569
570
571
572
573
574

575
576
577
578
579
580
581
		    (begin
		      (debug:print 0 *default-log-port* "ERROR: (tt-port ttdat) no port set! Exiting.")
		      (exit)))))
	  
	  ;; create a servinfo file start keep-running
	  (tt:create-server-registration-file ttdat dbfname)
	  (procinf-status-set! *procinf* "running")

	  (dbfile:with-no-sync-db
	   nosyncdbpath
	   (lambda (nsdb)
	     (dbfile:insert-or-update-process nsdb *procinf*)))
	  (thread-start! run-thread)

	  (thread-join! run-thread) ;; run thread will exit on timeout or other conditions







>







555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
		    (begin
		      (debug:print 0 *default-log-port* "ERROR: (tt-port ttdat) no port set! Exiting.")
		      (exit)))))
	  
	  ;; create a servinfo file start keep-running
	  (tt:create-server-registration-file ttdat dbfname)
	  (procinf-status-set! *procinf* "running")
	  (tt-state-set! ttdat 'running)
	  (dbfile:with-no-sync-db
	   nosyncdbpath
	   (lambda (nsdb)
	     (dbfile:insert-or-update-process nsdb *procinf*)))
	  (thread-start! run-thread)

	  (thread-join! run-thread) ;; run thread will exit on timeout or other conditions