Megatest

Diff
Login

Differences From Artifact [dce077e9b3]:

To Artifact [f86e80cfc1]:


325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
;;		  ;; seems like always)?
;;		  ((< (- (current-seconds)(runs:dat-beginning-of-time runsdat)) 30) ;; for the first 30 seconds do not throttle in any way
;;		   0)
;;        	  ((> (runs:dat-can-run-more-tests-count runsdat) 20) ;; original intent was - save cycles, wait a long time
;;		   (if (runs:lownoise "waiting on tasks" 60)(debug:print-info 2 *default-log-port* "waiting for tasks to complete, sleeping briefly ..."))
;;		   10)  ;; obviously haven't had any work to do for a while
;;		  (else 0)))
;;		   ;; if have a number for inter-test-delay, use it, else don't delay much, maybe even zero?
;;		   (configf:lookup-number *configdat* "setup" "inter-test-delay" default: 0.01)
;;		   )))
  
  (let* ((num-running             (rmt:get-count-tests-running run-id #f)) ;; fastmode=no
	 (num-running-in-jobgroup (rmt:get-count-tests-running-in-jobgroup run-id jobgroup))
	 (job-group-limit         (let ((jobg-count (configf:lookup *configdat* "jobgroups" jobgroup)))
				    (if (string? jobg-count)
					(string->number jobg-count)
					jobg-count))))







|
|
|







325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
;;		  ;; seems like always)?
;;		  ((< (- (current-seconds)(runs:dat-beginning-of-time runsdat)) 30) ;; for the first 30 seconds do not throttle in any way
;;		   0)
;;        	  ((> (runs:dat-can-run-more-tests-count runsdat) 20) ;; original intent was - save cycles, wait a long time
;;		   (if (runs:lownoise "waiting on tasks" 60)(debug:print-info 2 *default-log-port* "waiting for tasks to complete, sleeping briefly ..."))
;;		   10)  ;; obviously haven't had any work to do for a while
;;		  (else 0)))
;; 		   ;; if have a number for inter-test-delay, use it, else don't delay much, maybe even zero?
;; 		   (configf:lookup-number *configdat* "setup" "inter-test-delay" default: 0.01)
;; 		   )))
  
  (let* ((num-running             (rmt:get-count-tests-running run-id #f)) ;; fastmode=no
	 (num-running-in-jobgroup (rmt:get-count-tests-running-in-jobgroup run-id jobgroup))
	 (job-group-limit         (let ((jobg-count (configf:lookup *configdat* "jobgroups" jobgroup)))
				    (if (string? jobg-count)
					(string->number jobg-count)
					jobg-count))))
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
    ((and (null? fails) ;; have not-started tests, but unable to run them.  everything looks completed with no prospect of unsticking something that is stuck.  we should mark hed as moribund and exit or continue if there are more tests to consider
	   (null? prereq-fails)
	   (null? non-completed))
     (debug:print-info 4 *default-log-port* "cond branch - "  "ei-4")
      (if  (runs:can-keep-running? hed 20)
	  (begin
	    (runs:inc-cant-run-tests hed)
	    (debug:print-info 0 *default-log-port* "no fails in prerequisites for " hed " but also none running, keeping " hed " for now. Try count: " (hash-table-ref/default *seen-cant-run-tests* hed 0)) ;; 
	    ;; getting here likely means the system is way overloaded, kill a full minute before continuing
	    ;;
	    ;; No runsdat, can't do this yet
	    ;; (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat)))
	    ;;
	    (thread-sleep! 5) ;; TODO: gate by normalized server load > 1.0 (maxload config thing)
	    ;; num-retries code was here
	    ;; we use this opportunity to move contents of reg to tal
	    (list (car newtal)(append (cdr newtal) reg) '() reruns)) ;; an issue with prereqs not yet met?







|

|







992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
    ((and (null? fails) ;; have not-started tests, but unable to run them.  everything looks completed with no prospect of unsticking something that is stuck.  we should mark hed as moribund and exit or continue if there are more tests to consider
	   (null? prereq-fails)
	   (null? non-completed))
     (debug:print-info 4 *default-log-port* "cond branch - "  "ei-4")
      (if  (runs:can-keep-running? hed 20)
	  (begin
	    (runs:inc-cant-run-tests hed)
	    (debug:print-info 0 *default-log-port* "no fails in prerequisites for " hed " but also none running, keeping " hed " for now. Try count: " (hash-table-ref/default *seen-cant-run-tests* hed 0) ", going to wait 60 sec.") ;; 
	    ;; getting here likely means the system is way overloaded, kill a full minute before continuing
	    ;; (thread-sleep! 60) ;; TODO: gate by normalized server load > 1.0 (maxload config thing) CHECKTHIS!!!
	    ;; No runsdat, can't do this yet
	    ;; (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat)))
	    ;;
	    (thread-sleep! 5) ;; TODO: gate by normalized server load > 1.0 (maxload config thing)
	    ;; num-retries code was here
	    ;; we use this opportunity to move contents of reg to tal
	    (list (car newtal)(append (cdr newtal) reg) '() reruns)) ;; an issue with prereqs not yet met?