@@ -55,73 +55,75 @@ ;; ;; lookup_server, need to remove *runremote* stuff ;; (define (client:setup run-id #!key (remaining-tries 100) (failed-connects 0)) (debug:print-info 2 "client:setup remaining-tries=" remaining-tries) - (if (<= remaining-tries 0) - (begin - (debug:print 0 "ERROR: failed to start or connect to server for run-id " run-id) - (exit 1)) - (let ((host-info (hash-table-ref/default *runremote* run-id #f))) - (if host-info - (let* ((iface (http-transport:server-dat-get-iface host-info)) - (port (http-transport:server-dat-get-port host-info)) - (start-res (http-transport:client-connect iface port)) - (ping-res (rmt:login-no-auto-client-setup start-res run-id))) - (if ping-res ;; sucessful login? - (begin - (debug:print-info 2 "client:setup, ping is good using host-info=" host-info ", remaining-tries=" remaining-tries) - ;; Why add the close-connections here? - ;; (http-transport:close-connections run-id) - (hash-table-set! *runremote* run-id start-res) - start-res) ;; return the server info - ;; have host info but no ping. shutdown the current connection and try again - (begin ;; login failed - (debug:print-info 1 "client:setup, ping is bad for start-res=" start-res " and *runremote*=" host-info) - (http-transport:close-connections run-id) - (hash-table-delete! *runremote* run-id) - (if (< remaining-tries 8) - (thread-sleep! 5) - (thread-sleep! 1)) - (client:setup run-id remaining-tries: (- remaining-tries 1))))) - ;; YUK: rename server-dat here - (let* ((server-dat (tasks:get-server (tasks:get-db) run-id))) - (debug:print-info 4 "client:setup server-dat=" server-dat ", remaining-tries=" remaining-tries) - (if server-dat - (let* ((iface (tasks:hostinfo-get-interface server-dat)) - (port (tasks:hostinfo-get-port server-dat)) - (start-res (http-transport:client-connect iface port)) - (ping-res (rmt:login-no-auto-client-setup start-res run-id))) - (if (and start-res - ping-res) - (begin - (hash-table-set! *runremote* run-id start-res) - (debug:print-info 2 "connected to " (http-transport:server-dat-make-url start-res)) - start-res) - (begin ;; login failed but have a server record, clean out the record and try again - (debug:print-info 0 "client:setup, login failed, will attempt to start server ... start-res=" start-res ", run-id=" run-id ", server-dat=" server-dat) - (http-transport:close-connections run-id) - (hash-table-delete! *runremote* run-id) - (tasks:server-force-clean-run-record (tasks:get-db) - run-id - (tasks:hostinfo-get-interface server-dat) - (tasks:hostinfo-get-port server-dat) - " client:setup (server-dat = #t)") - (thread-sleep! 2) - (server:try-running run-id) - (thread-sleep! 10) ;; give server a little time to start up - (client:setup run-id remaining-tries: (- remaining-tries 1))))) - (begin ;; no server registered - (let ((num-available (tasks:num-in-available-state (tasks:get-db) run-id))) - (debug:print-info 0 "client:setup, no server registered, remaining-tries=" remaining-tries " num-available=" num-available) - (thread-sleep! 2) - (if (< num-available 2) - (begin - ;; (open-run-close tasks:server-clean-out-old-records-for-run-id tasks:open-db run-id " client:setup (server-dat=#f)") - (server:try-running run-id))) - (thread-sleep! 10) ;; give server a little time to start up - (client:setup run-id remaining-tries: (- remaining-tries 1)))))))))) + (let* ((tdbdat (tasks:open-db)) + (tdb (db:dbdat-get-db tdbdat))) + (if (<= remaining-tries 0) + (begin + (debug:print 0 "ERROR: failed to start or connect to server for run-id " run-id) + (exit 1)) + (let ((host-info (hash-table-ref/default *runremote* run-id #f))) + (if host-info + (let* ((iface (http-transport:server-dat-get-iface host-info)) + (port (http-transport:server-dat-get-port host-info)) + (start-res (http-transport:client-connect iface port)) + (ping-res (rmt:login-no-auto-client-setup start-res run-id))) + (if ping-res ;; sucessful login? + (begin + (debug:print-info 2 "client:setup, ping is good using host-info=" host-info ", remaining-tries=" remaining-tries) + ;; Why add the close-connections here? + ;; (http-transport:close-connections run-id) + (hash-table-set! *runremote* run-id start-res) + start-res) ;; return the server info + ;; have host info but no ping. shutdown the current connection and try again + (begin ;; login failed + (debug:print-info 1 "client:setup, ping is bad for start-res=" start-res " and *runremote*=" host-info) + (http-transport:close-connections run-id) + (hash-table-delete! *runremote* run-id) + (if (< remaining-tries 8) + (thread-sleep! 5) + (thread-sleep! 1)) + (client:setup run-id remaining-tries: (- remaining-tries 1))))) + ;; YUK: rename server-dat here + (let* ((server-dat (tasks:get-server (db:delay-if-busy tdbdat) run-id))) + (debug:print-info 4 "client:setup server-dat=" server-dat ", remaining-tries=" remaining-tries) + (if server-dat + (let* ((iface (tasks:hostinfo-get-interface server-dat)) + (port (tasks:hostinfo-get-port server-dat)) + (start-res (http-transport:client-connect iface port)) + (ping-res (rmt:login-no-auto-client-setup start-res run-id))) + (if (and start-res + ping-res) + (begin + (hash-table-set! *runremote* run-id start-res) + (debug:print-info 2 "connected to " (http-transport:server-dat-make-url start-res)) + start-res) + (begin ;; login failed but have a server record, clean out the record and try again + (debug:print-info 0 "client:setup, login failed, will attempt to start server ... start-res=" start-res ", run-id=" run-id ", server-dat=" server-dat) + (http-transport:close-connections run-id) + (hash-table-delete! *runremote* run-id) + (tasks:server-force-clean-run-record (db:delay-if-busy tdbdat) + run-id + (tasks:hostinfo-get-interface server-dat) + (tasks:hostinfo-get-port server-dat) + " client:setup (server-dat = #t)") + (thread-sleep! 2) + (server:try-running run-id) + (thread-sleep! 10) ;; give server a little time to start up + (client:setup run-id remaining-tries: (- remaining-tries 1))))) + (begin ;; no server registered + (let ((num-available (tasks:num-in-available-state (db:dbdat-get-db tdbdat) run-id))) + (debug:print-info 0 "client:setup, no server registered, remaining-tries=" remaining-tries " num-available=" num-available) + (thread-sleep! 2) + (if (< num-available 2) + (begin + ;; (open-run-close tasks:server-clean-out-old-records-for-run-id tasks:open-db run-id " client:setup (server-dat=#f)") + (server:try-running run-id))) + (thread-sleep! 10) ;; give server a little time to start up + (client:setup run-id remaining-tries: (- remaining-tries 1))))))))))) ;; keep this as a function to ease future (define (client:start run-id server-info) (http-transport:client-connect (tasks:hostinfo-get-interface server-info) (tasks:hostinfo-get-port server-info)))