@@ -1265,12 +1265,12 @@ (begin ;; couldn't run, take a breather (if (runs:lownoise "Waiting for more work to do..." 60) (debug:print-info 0 *default-log-port* "Waiting for more work to do...")) - (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat))) - (thread-sleep! 1) + ;; (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat))) + (thread-sleep! 5) (list (car newtal)(cdr newtal) reg reruns)) ;; the waiton is FAIL so no point in trying to run hed ever again (begin (let ((my-test-id (rmt:get-test-id run-id test-name item-path))) (mt:test-set-state-status-by-id-unless-completed run-id my-test-id "COMPLETED" "PREQ_FAIL" "Failed to run due to failed prerequisites2")) @@ -1800,11 +1800,11 @@ ;; this is the point where everything is launched and now you can mark the run in metadata table as all launched (rmt:set-var (conc "lunch-complete-" run-id) "yes") ;; now *if* -run-wait we wait for all tests to be done ;; Now wait for any RUNNING tests to complete (if in run-wait mode) - (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat))) + ;; (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat))) (thread-sleep! 10) ;; I think there is a race condition here. Let states/statuses settle (let wait-loop ((num-running (rmt:get-count-tests-running-for-run-id run-id #t)) ;; fastmode=yes (prev-num-running 0)) ;; (debug:print-info 13 *default-log-port* "num-running=" num-running ", prev-num-running=" prev-num-running) @@ -1821,12 +1821,12 @@ (set! last-time-incomplete (current-seconds)) ;; FIXME, this might be causing slow down - use of set! (rmt:find-and-mark-incomplete run-id #f) (debug:print-info 0 *default-log-port* "run-wait specified, waiting on " actual-num-running " tests in RUNNING, REMOTEHOSTSTART or LAUNCHED state at " (time->string (seconds->local-time (current-seconds)))))) - (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat))) - (thread-sleep! 1) ;; (if (>= num-running max-concurrent-jobs) 5 1)) + ;; (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat))) + (thread-sleep! 5) ;; (if (>= num-running max-concurrent-jobs) 5 1)) (wait-loop (rmt:get-count-tests-running-for-run-id run-id #t) ;; fastmode=yes num-running)))) ;; LET* ((test-record ;; we get here on "drop through". All done! ;; this is moved to runs:run-testes since this function is getting called twice to ensure everthing is completed. @@ -1959,13 +1959,13 @@ (set! test-id (rmt:get-test-id run-id test-name item-path)))) (debug:print-info 4 *default-log-port* "test-id=" test-id ", run-id=" run-id ", test-name=" test-name ", item-path=\"" item-path "\"") (set! testdat (rmt:get-test-info-by-id run-id test-id)) (if (not testdat) (begin - (debug:print-info 0 *default-log-port* "WARNING: server is overloaded, trying again in one second") - (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat))) - (thread-sleep! 1) + (debug:print-info 0 *default-log-port* "WARNING: server is overloaded, trying again in two seconds") + ;; (if (runs:dat-load-mgmt-function runsdat)((runs:dat-load-mgmt-function runsdat))) + (thread-sleep! 2) (loop))))) (if (not testdat) ;; should NOT happen (debug:print-error 0 *default-log-port* "failed to get test record for test-id " test-id)) (set! test-id (db:test-get-id testdat)) (if (common:file-exists? test-path)