@@ -381,10 +381,11 @@
     (tests:set-full-meta-info #f test-id run-id (calc-minutes) work-area 10)
     (let loop ((minutes   (calc-minutes))
 	       (cpu-load  (alist-ref 'adj-core-load (common:get-normalized-cpu-load #f)))
 	       (disk-free (get-df (current-directory)))
                (last-sync (current-seconds)))
+      (print "launch:monitor-job - top of loop encountered at "(current-seconds)" with last-sync="last-sync)
       (let* ((over-time     (> (current-seconds) (+ last-sync update-period)))
              (new-cpu-load  (let* ((load  (alist-ref 'adj-core-load (common:get-normalized-cpu-load #f)))
                                    (delta (abs (- load cpu-load))))
                               (if (> delta 0.1) ;; don't bother updating with small changes
                                   load
@@ -402,26 +403,30 @@
              (test-info   (rmt:get-test-info-by-id run-id test-id))
              (state       (db:test-get-state test-info))
              (status      (db:test-get-status test-info))
              (kill-reason  "no kill reason specified")
              (kill-job?    #f))
-
+        (print "launch:monitor-job - decision time encountered at "(current-seconds)" with last-sync="last-sync" do-sync="do-sync" over-time="over-time" update-period="update-period)
         (cond
          ((test-get-kill-request run-id test-id)
           (set! kill-reason "KILLING TEST since received kill request (KILLREQ)")
           (set! kill-job? #t))
          ((and runtlim (> (- (current-seconds) start-seconds) runtlim))
           (set! kill-reason (conc "KILLING TEST DUE TO TIME LIMIT EXCEEDED! Runtime=" run-seconds " seconds, limit=" runtlim))
           (set! kill-job? #t))
          ((equal? status "DEAD")
-          (set! kill-reason "KILLING TEST because it was marked as DEAD by launch:handle-zombie-tests")
+          (set! kill-reason "KILLING TEST because it was marked as DEAD by launch:handle-zombie-tests (might indicate really overloaded server or else overzealous setup.deadtime)")
           (set! kill-job? #t)))
 
         (debug:print 4 *default-log-port* "cpu: " new-cpu-load " disk: " new-disk-free " last-sync: " last-sync " do-sync: " do-sync)
         (launch:handle-zombie-tests run-id)
-        (if do-sync
-            (tests:update-central-meta-info run-id test-id new-cpu-load new-disk-free (calc-minutes) #f #f))
+        (when do-sync
+          ;;(with-output-to-file (conc (getenv "MT_TEST_RUN_DIR") "/last-loadinfo.log" #:append)
+          ;;  (lambda () (pp (list (current-seconds) new-cpu-load new-disk-free (calc-minutes)))))
+          (print "launch:monitor-job - dosync started at "(current-seconds))
+          (tests:update-central-meta-info run-id test-id new-cpu-load new-disk-free (calc-minutes) #f #f)
+          (print "launch:monitor-job - dosync finished at "(current-seconds)))
         
 	(if kill-job? 
 	    (begin
               (debug:print-info 0 *default-log-port* "proceeding to kill test: "kill-reason)
 	      (mutex-lock! m)