Megatest

Diff
Login

Differences From Artifact [96c70e902e]:

To Artifact [f4c57969ab]:


565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580


581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601

602
603
604
605
606
607
608
;;======================================================================
;; END NEW SERVER METHOD
;;======================================================================

;; 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) 
  ;; 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* ((run-id            (let ((rid (args:get-arg "-run-id")))
			      (if rid
				  (string->number rid)
				  #f)))


	 (db-file           (db:run-id->path *toppath* run-id))
	 (sdat              #f)
	 (tmp-area          (common:get-db-tmp-area))
	 (server-start-time (current-seconds))
	 (pkts-dir          (get-pkts-dir))
	 (server-key        (server:mk-signature))
	 (server-info (let loop ((start-time (current-seconds))
				 (changed    #t)
				 (last-sdat  "not this"))
                        (begin ;; let ((sdat #f))
			  (thread-sleep! 0.01)
			  (debug:print-info 0 *default-log-port* "Waiting for server alive signature")
                          (mutex-lock! *heartbeat-mutex*)
                          (set! sdat *server-info*)
                          (mutex-unlock! *heartbeat-mutex*)
                          (if (and sdat
				   (not changed)
				   (> (- (current-seconds) start-time) 2))
			      (begin
				(debug:print-info 0 *default-log-port* "Received server alive signature, now attempting to lock in server")
				;; create a server pkt in *toppath*/.meta/srvpkts

				(register-server pkts-dir *srvpktspec* (get-host-name)
						 (cadr sdat) server-key (car sdat) db-file)

				;; now read pkts and see if we are a contender
				(let* ((all-pkts     (get-all-server-pkts pkts-dir *srvpktspec*))
				       (viables      (get-viable-servers all-pkts db-file))
				       (best-srv     (get-best-candidate viables db-file))







|




|



>
>
|

|


















>







565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
;;======================================================================
;; END NEW SERVER METHOD
;;======================================================================

;; 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 dbname) 
  ;; 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* ((run-id            (let ((rid (args:get-arg "-run-id"))) ;; consider getting rid of the -run-id mechanism
			      (if rid
				  (string->number rid)
				  #f)))
	 (db-file           (if dbname
				(db:dbname->path *toppath* dbname)
				(db:run-id->path *toppath* run-id)))
	 (sdat              #f)
	 ;; (tmp-area          (common:get-db-tmp-area))
	 (server-start-time (current-seconds))
	 (pkts-dir          (get-pkts-dir))
	 (server-key        (server:mk-signature))
	 (server-info (let loop ((start-time (current-seconds))
				 (changed    #t)
				 (last-sdat  "not this"))
                        (begin ;; let ((sdat #f))
			  (thread-sleep! 0.01)
			  (debug:print-info 0 *default-log-port* "Waiting for server alive signature")
                          (mutex-lock! *heartbeat-mutex*)
                          (set! sdat *server-info*)
                          (mutex-unlock! *heartbeat-mutex*)
                          (if (and sdat
				   (not changed)
				   (> (- (current-seconds) start-time) 2))
			      (begin
				(debug:print-info 0 *default-log-port* "Received server alive signature, now attempting to lock in server")
				;; create a server pkt in *toppath*/.meta/srvpkts
				
				(register-server pkts-dir *srvpktspec* (get-host-name)
						 (cadr sdat) server-key (car sdat) db-file)

				;; now read pkts and see if we are a contender
				(let* ((all-pkts     (get-all-server-pkts pkts-dir *srvpktspec*))
				       (viables      (get-viable-servers all-pkts db-file))
				       (best-srv     (get-best-candidate viables db-file))
641
642
643
644
645
646
647

648

649
650
651
652
653
654
655
	       (server-state 'available)
	       (bad-sync-count 0)
	       (start-time     (current-milliseconds)))
      ;; Use this opportunity to sync the tmp db to megatest.db
      (if (not *dbstruct-db* )
	  (let ((watchdog (bdat-watchdog *bdat*)))
	    (debug:print 0 *default-log-port* "SERVER: dbprep")

	    (db:setup run-id) ;; sets *dbstruct-db* as side effect

	    (debug:print 0 *default-log-port* "SERVER: running, megatest version: " (common:get-full-version)) ;; NOTE: the server is NOT yet marked as running in the log. We do that in the keep-running routine.
	    (if watchdog
		(if (not (member (thread-state watchdog) '(ready running blocked sleeping dead)))
		    (begin
		      (debug:print-info 0 "Starting watchdog thread (in state "(thread-state watchdog)")")
		      (thread-start! watchdog)))
		(debug:print 0 *default-log-port* "ERROR: *watchdog* not setup, cannot start it."))))







>
|
>







644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
	       (server-state 'available)
	       (bad-sync-count 0)
	       (start-time     (current-milliseconds)))
      ;; Use this opportunity to sync the tmp db to megatest.db
      (if (not *dbstruct-db* )
	  (let ((watchdog (bdat-watchdog *bdat*)))
	    (debug:print 0 *default-log-port* "SERVER: dbprep")
	    
	    (db:setup dbname) ;; sets *dbstruct-db* as side effect

	    (debug:print 0 *default-log-port* "SERVER: running, megatest version: " (common:get-full-version)) ;; NOTE: the server is NOT yet marked as running in the log. We do that in the keep-running routine.
	    (if watchdog
		(if (not (member (thread-state watchdog) '(ready running blocked sleeping dead)))
		    (begin
		      (debug:print-info 0 "Starting watchdog thread (in state "(thread-state watchdog)")")
		      (thread-start! watchdog)))
		(debug:print 0 *default-log-port* "ERROR: *watchdog* not setup, cannot start it."))))
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
;; Call this to start the actual server
;;

;; all routes though here end in exit ...
;;
;; start_server? 
;;
(define (http-transport:launch)
  (let* ((tmp-area            (common:get-db-tmp-area))
	 (server-start        (conc tmp-area "/.server-start"))
	 (server-started      (conc tmp-area "/.server-started"))
	 (start-time          (common:lazy-modification-time server-start))
	 (started-time        (common:lazy-modification-time server-started))
	 (server-starting     (< start-time started-time)) ;; if start-time is less than started-time then a server is still starting
	 (start-time-old      (> (- (current-seconds) start-time) 5))
         (cleanup-proc        (lambda (msg)
                                (let* ((serv-fname      (conc "server-" (current-process-id) "-" (get-host-name) ".log"))
                                       (full-serv-fname (conc *toppath* "/logs/" serv-fname))
                                       (new-serv-fname  (conc *toppath* "/logs/" "defunct-" serv-fname)))
                                  (debug:print 0 *default-log-port* msg)
                                  (if (common:file-exists? full-serv-fname)
                                      (system (conc "sleep 1;mv -f " full-serv-fname " " new-serv-fname))
                                      (debug:print 0 *default-log-port* "INFO: cannot move " full-serv-fname " to " new-serv-fname))
                                  (exit)))))
    (common:save-pkt `((action . start)
		       (T      . server)
		       (pid    . ,(current-process-id)))
		     *configdat* #t)
    (let* ((th2 (make-thread (lambda ()
                               (debug:print-info 0 *default-log-port* "Server run thread started")
                               (http-transport:run 
                                (if (args:get-arg "-server")
                                    (args:get-arg "-server")
                                    "-")
                                )) "Server run"))
           (th3 (make-thread (lambda ()
                               (debug:print-info 0 *default-log-port* "Server monitor thread started")
                               (http-transport:keep-running)
                               "Keep running"))))
      (thread-start! th2)
      (thread-sleep! 0.252) ;; give the server time to settle before starting the keep-running monitor.
      (thread-start! th3)
      (set! *didsomething* #t)
      (thread-join! th2)
      (exit))))







|
|
|
|
|
|
|
|









|












|







759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
;; Call this to start the actual server
;;

;; all routes though here end in exit ...
;;
;; start_server? 
;;
(define (http-transport:launch dbname)
  (let* (;; (tmp-area            (common:get-db-tmp-area))
	 ;; (server-start        (conc tmp-area "/.server-start"))
	 ;; (server-started      (conc tmp-area "/.server-started"))
	 ;; (start-time          (common:lazy-modification-time server-start))
	 ;; (started-time        (common:lazy-modification-time server-started))
	 ;; (server-starting     (< start-time started-time)) ;; if start-time is less than started-time then a server is still starting
	 ;; (start-time-old      (> (- (current-seconds) start-time) 5))
         (cleanup-proc        (lambda (msg)
                                (let* ((serv-fname      (conc "server-" (current-process-id) "-" (get-host-name) ".log"))
                                       (full-serv-fname (conc *toppath* "/logs/" serv-fname))
                                       (new-serv-fname  (conc *toppath* "/logs/" "defunct-" serv-fname)))
                                  (debug:print 0 *default-log-port* msg)
                                  (if (common:file-exists? full-serv-fname)
                                      (system (conc "sleep 1;mv -f " full-serv-fname " " new-serv-fname))
                                      (debug:print 0 *default-log-port* "INFO: cannot move " full-serv-fname " to " new-serv-fname))
                                  (exit)))))
    #;(common:save-pkt `((action . start)
		       (T      . server)
		       (pid    . ,(current-process-id)))
		     *configdat* #t)
    (let* ((th2 (make-thread (lambda ()
                               (debug:print-info 0 *default-log-port* "Server run thread started")
                               (http-transport:run 
                                (if (args:get-arg "-server")
                                    (args:get-arg "-server")
                                    "-")
                                )) "Server run"))
           (th3 (make-thread (lambda ()
                               (debug:print-info 0 *default-log-port* "Server monitor thread started")
                               (http-transport:keep-running dbname)
                               "Keep running"))))
      (thread-start! th2)
      (thread-sleep! 0.252) ;; give the server time to settle before starting the keep-running monitor.
      (thread-start! th3)
      (set! *didsomething* #t)
      (thread-join! th2)
      (exit))))
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
;; Call this to start the actual server
;;

;; all routes though here end in exit ...
;;
;; start_server
;;
(define (server:launch run-id transport-type)
    (http-transport:launch))

;;======================================================================
;; S E R V E R   U T I L I T I E S 
;;======================================================================

;; NOT USED (well, ok, reference in rpc-transport but otherwise not used).
;;







|
|







859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
;; Call this to start the actual server
;;

;; all routes though here end in exit ...
;;
;; start_server
;;
(define (server:launch dbname)
    (http-transport:launch dbname))

;;======================================================================
;; S E R V E R   U T I L I T I E S 
;;======================================================================

;; NOT USED (well, ok, reference in rpc-transport but otherwise not used).
;;