Megatest

    ((zmq)  (zmq:client-connect  iface port))
    (else
     (debug:print 0 *default-log-port* "ERROR: transport " (remote-transport *runremote*) " not supported (5)")
     (exit))))

(define (client:setup  run-id #!key (remaining-tries 10) (failed-connects 0))
  (case (server:get-transport)
    ((rpc) (let ((res (rpc-transport:client-setup run-id remaining-tries: remaining-tries failed-connects: failed-connects)))
             (remote-conndat-set! *runremote* runremote-server-dat)
             res))
    ((http)(client:setup-http run-id remaining-tries: remaining-tries failed-connects: failed-connects))
    (else
     (debug:print 0 *default-log-port* "ERROR: transport " (remote-transport *runremote*) " not supported (6)")
     (exit)))) ;; (client:setup-rpc run-id))))

;; (define (client:login-no-auto-setup server-info run-id)
;;   (case (server:get-transport)

;;   2. We are a run tests, list runs or other interactive process and we must figure out
;;      *transport-type* and *runremote* from the monitor.db
;;
;; client:setup
;;
;; lookup_server, need to remove *runremote* stuff
;;

(define (client:setup-rpc run-id #!key (remaining-tries 10) (failed-connects 0))
  (debug:print-info 2 *default-log-port* "client:setup-rpc remaining-tries=" remaining-tries)
  (let* ((server-dat (tasks:get-server (db:delay-if-busy (tasks:open-db)) run-id))
         (num-available (tasks:num-in-available-state (db:delay-if-busy (tasks:open-db)) run-id)))
    (cond
     ((<= remaining-tries 0)
      (debug:print-error 0 *default-log-port* "failed to start or connect to server for run-id " run-id)
      (exit 1))
     (server-dat
      (debug:print-info 4 *default-log-port* "client:setup-rpc server-dat=" server-dat ", remaining-tries=" remaining-tries)

      (rpc-transport:client-setup run-id server-dat remaining-tries: remaining-tries))
     (else
      (if (< num-available 2)
          (server:try-running run-id))
      (thread-sleep! (+ 2 (random (- 20 remaining-tries)))) ;; give server a little time to start up, randomize a little to avoid start storms.
      (client:setup run-id remaining-tries: (- remaining-tries 1))))))


(define (client:setup-http run-id #!key (remaining-tries 10) (failed-connects 0))
  (debug:print-info 2 *default-log-port* "client:setup remaining-tries=" remaining-tries)
  (let* ((tdbdat (tasks:open-db)))
    (if (<= remaining-tries 0)
	(begin
	  (debug:print-error 0 *default-log-port* "failed to start or connect to server for run-id " run-id)
	  (exit 1))

;;
;; connect
;;
(define (http-transport:client-connect iface port)
  (let* ((api-url      (conc "http://" iface ":" port "/api"))
	 (api-uri      (uri-reference (conc "http://" iface ":" port "/api")))
	 (api-req      (make-request method: 'POST uri: api-uri))
	 (server-dat   (vector iface port api-uri api-url api-req (current-seconds) 'http)))
    server-dat))

;; 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 server-id run-id)
  ;; if none running or if > 20 seconds since 

    (set! *last-db-access* (current-seconds)) ;; bump *last-db-access*; this will renew keep-running thread's lease on life for another (server:get-timeout) seconds
    ;;(BB> "in api-exec; last-db-access updated to "*last-db-access*)
    (mutex-unlock! *heartbeat-mutex*)

    res))











;; retry an operation (depends on srfi-18)
;; ==================
;; idea here is to avoid spending time on coding retrying something.  Trying to be generic here.
;;
;; Exception handling:
;; -------------------
;; if evaluating the thunk results in exception, it will be retried.

          #t)
        (begin
          (BB> "Self test fail.  login-res="login-res" testing-res="testing-res" *toppath*="*toppath*)
           
          #f))
    res))

(define (rpc-transport:client-setup run-id server-dat #!key (remaining-tries 10))
  ;;(BB> "entered rpc-transport:client-setup with run-id="run-id" and server-dat="server-dat" and retries="remaining-tries)
  (tcp-buffer-size 0)
  (debug:print-info 0 *default-log-port* "rpc-transport:client-setup run-id="run-id" server-dat=" server-dat ", remaining-tries=" remaining-tries)
  (let* ((iface     (tasks:hostinfo-get-interface server-dat))
         (hostname  (tasks:hostinfo-get-hostname  server-dat))
         (port      (tasks:hostinfo-get-port      server-dat))
         (runremote-server-dat (vector iface port #f #f #f (current-seconds) 'rpc)) ;; http version := (vector iface port api-uri api-url api-req (current-seconds) 'http  )
         (ping-res (retry-thunk (lambda ()  ;; make 3 attempts to ping.
                                  ((rpc:procedure 'server:login iface port) *toppath*))
                                chatty: #f
                                retries: 3)))
    ;; we got here from rmt:get-connection-info on the condition that *runremote* has no entry for run-id...
    (if ping-res
        (begin
          (debug:print-info 0 *default-log-port* "rpc-transport:client-setup CONNECTION ESTABLISHED run-id="run-id" server-dat=" server-dat)

          runremote-server-dat)
        (begin ;; login failed but have a server record, clean out the record and try again
          (debug:print-info 0 *default-log-port* "rpc-transport:client-setup UNABLE TO CONNECT run-id="run-id" server-dat=" server-dat)
          (tasks:kill-server-run-id run-id)
          (tasks:server-force-clean-run-record  (db:delay-if-busy (tasks:open-db)) run-id iface port
                                                   " rpc-transport:client-setup (server-dat = #t)")
          (if (> remaining-tries 2)
              (thread-sleep! (+ 1 (random 5))) ;; spread out the starts a little
              (thread-sleep! (+ 15 (random 20)))) ;; it isn't going well. give it plenty of time
          (server:try-running run-id)
          (thread-sleep! 5)   ;; give server a little time to start up
          (client:setup run-id remaining-tries: (sub1 remaining-tries))))))