84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
|
;;======================================================================
;; S E R V E R U T I L I T I E S
;;======================================================================
;;======================================================================
;; C L I E N T S
;;======================================================================
(define *http-mutex* (make-mutex))
;; NOTE: Large block of code from 32436b426188080f72fceb6894af541fbad9921e removed here
;; I'm pretty sure it is defunct.
;; This next block all imported en-mass from the api branch
(define *http-requests-in-progress* 0)
(define *http-connections-next-cleanup* (current-seconds))
(define (http-transport:get-time-to-cleanup)
(let ((res #f))
(mutex-lock! *http-mutex*)
(set! res (> (current-seconds) *http-connections-next-cleanup*))
(mutex-unlock! *http-mutex*)
res))
(define (http-transport:inc-requests-count)
(mutex-lock! *http-mutex*)
(set! *http-requests-in-progress* (+ 1 *http-requests-in-progress*))
;; Use this opportunity to slow things down iff there are too many requests in flight
(if (> *http-requests-in-progress* 5)
(begin
(debug:print-info 0 *default-log-port* "Whoa there buddy, ease up...")
(thread-sleep! 1)))
(mutex-unlock! *http-mutex*))
(define (http-transport:dec-requests-count proc)
(mutex-lock! *http-mutex*)
(proc)
(set! *http-requests-in-progress* (- *http-requests-in-progress* 1))
(mutex-unlock! *http-mutex*))
(define (http-transport:dec-requests-count-and-close-all-connections)
(set! *http-requests-in-progress* (- *http-requests-in-progress* 1))
(let loop ((etime (+ (current-seconds) 5))) ;; give up in five seconds
(if (> *http-requests-in-progress* 0)
(if (> etime (current-seconds))
(begin
(thread-sleep! 0.05)
(loop etime))
(debug:print-error 0 *default-log-port* "requests still in progress after 5 seconds of waiting. I'm going to pass on cleaning up http connections"))
(close-all-connections!)))
(set! *http-connections-next-cleanup* (+ (current-seconds) 10))
(mutex-unlock! *http-mutex*))
(define (http-transport:inc-requests-and-prep-to-close-all-connections)
(mutex-lock! *http-mutex*)
(set! *http-requests-in-progress* (+ 1 *http-requests-in-progress*)))
;; Send "cmd" with json payload "params" to serverdat and receive result
;;
(define (http-transport:client-api-send-receive run-id serverdat cmd params #!key (numretries 3)(area-dat #f))
(let* ((fullurl (if (vector? serverdat)
(http-transport:server-dat-get-api-req serverdat)
(begin
(debug:print 0 *default-log-port* "FATAL ERROR: http-transport:client-api-send-receive called with no server info")
(exit 1))))
(res (vector #f "uninitialized"))
(success #t)
(sparams (db:obj->string params transport: 'http))
(runremote (or area-dat *runremote*))
(server-id (if (vector? serverdat)
(http-transport:server-dat-get-server-id serverdat)
(begin
(debug:print 0 *default-log-port* "FATAL ERROR: http-transport:client-api-send-receive called with no server info")
(exit 1)))))
(debug:print-info 11 *default-log-port* "cmd=" cmd " fullurl=" fullurl " server-id=" server-id " current time:" (current-seconds))
;; set up the http-client here
(max-retry-attempts 1)
;; consider all requests indempotent
(retry-request? (lambda (request)
#f))
;; send the data and get the response
;; extract the needed info from the http data and
;; process and return it.
(let* ((send-recieve (lambda ()
(mutex-lock! *http-mutex*)
;; (condition-case (with-input-from-request "http://localhost"; #f read-lines)
;; ((exn http client-error) e (print e)))
(set! res (vector ;;; DON'T FORGET - THIS IS THE CLIENT SIDE! NOTE: consider moving this to client.scm since we are only supporting http transport at this time.
success
(db:string->obj
(handle-exceptions
exn
(let ((call-chain (get-call-chain))
(msg ((condition-property-accessor 'exn 'message) exn)))
(set! success #f)
(if (debug:debug-mode 1)
(debug:print-info 0 *default-log-port* "couldn't talk to server, trying again ...")
(begin
(debug:print 0 *default-log-port* "WARNING: failure in with-input-from-request to " fullurl ".")
(debug:print 0 *default-log-port* " message: " msg ", exn=" exn)
(debug:print 0 *default-log-port* " cmd: " cmd " params: " params " key:" (or server-id "thekey"))
(debug:print 0 *default-log-port* " call-chain: " call-chain)))
(if runremote
(remote-conndat-set! runremote #f))
;; Killing associated server to allow clean retry.")
;; (tasks:kill-server-run-id run-id) ;; better to kill the server in the logic that called this routine?
(mutex-unlock! *http-mutex*)
;;; (signal (make-composite-condition
;;; (make-property-condition 'commfail 'message "failed to connect to server")))
;;; "communications failed"
(db:obj->string #f))
(with-input-from-request ;; was dat
fullurl
(list (cons 'key (or server-id "thekey"))
(cons 'cmd cmd)
(cons 'params sparams))
read-string))
transport: 'http)
0)) ;; added this speculatively
;; Shouldn't this be a call to the managed call-all-connections stuff above?
(close-all-connections!)
(mutex-unlock! *http-mutex*)
))
(time-out (lambda ()
(thread-sleep! 45)
(debug:print 0 *default-log-port* "WARNING: send-receive took more than 45 seconds!!")
#f))
(th1 (make-thread send-recieve "with-input-from-request"))
(th2 (make-thread time-out "time out")))
(thread-start! th1)
(thread-start! th2)
(thread-join! th1)
(vector-set! res 0 success)
(thread-terminate! th2)
(if (vector? res)
(if (vector-ref res 0) ;; this is the first flag or the second flag?
(let* ((res-dat (vector-ref res 1)))
(if (and (string? res-dat) (string-contains res-dat "server-id mismatch"))
(signal (make-composite-condition
(make-property-condition
'servermismatch
'message (vector-ref res 1))))
res)) ;; this is the *inner* vector? seriously? why?
(if (debug:debug-mode 11)
(let ((call-chain (get-call-chain))) ;; note: this code also called in nmsg-transport - consider consolidating it
(print-call-chain (current-error-port))
(debug:print-error 11 *default-log-port* "error above occured at server, res=" res) ;; " message: " ((condition-property-accessor 'exn 'message) exn))
(debug:print 11 *default-log-port* " server call chain:")
(pp (vector-ref res 1) (current-error-port))
(signal (vector-ref res 0)))
res))
(signal (make-composite-condition
(make-property-condition
'timeout
'message "nmsg-transport:client-api-send-receive-raw timed out talking to server")))))))
;; careful closing of connections stored in *runremote*
;;
(define (http-transport:close-connections #!key (area-dat #f))
(let* ((runremote (or area-dat *runremote*))
(server-dat (if runremote
(remote-conndat runremote)
#f))) ;; (hash-table-ref/default *runremote* run-id #f)))
(if (vector? server-dat)
(let ((api-dat (http-transport:server-dat-get-api-uri server-dat)))
(handle-exceptions
exn
(begin
(print-call-chain *default-log-port*)
(debug:print-error 0 *default-log-port* " closing connection failed with error: " ((condition-property-accessor 'exn 'message) exn) ", exn=" exn))
(close-connection! api-dat)
;;(close-idle-connections!)
#t))
#f)))
(define (make-http-transport:server-dat)(make-vector 6))
(define (http-transport:server-dat-get-iface vec) (vector-ref vec 0))
(define (http-transport:server-dat-get-port vec) (vector-ref vec 1))
(define (http-transport:server-dat-get-api-uri vec) (vector-ref vec 2))
(define (http-transport:server-dat-get-api-url vec) (vector-ref vec 3))
(define (http-transport:server-dat-get-api-req vec) (vector-ref vec 4))
(define (http-transport:server-dat-get-last-access vec) (vector-ref vec 5))
;(define (http-transport:server-dat-get-socket vec) (vector-ref vec 6))
(define (http-transport:server-dat-get-server-id vec) (vector-ref vec 6))
(define (http-transport:server-dat-make-url vec)
(if (and (http-transport:server-dat-get-iface vec)
(http-transport:server-dat-get-port vec))
(conc "http://"
(http-transport:server-dat-get-iface vec)
":"
(http-transport:server-dat-get-port vec))
#f))
(define (http-transport:server-dat-update-last-access vec)
(if (vector? vec)
(vector-set! vec 5 (current-seconds))
(begin
(print-call-chain (current-error-port))
(debug:print-error 0 *default-log-port* "call to http-transport:server-dat-update-last-access with non-vector!!"))))
;;
;; connect
;;
(define (http-transport:client-connect iface port server-id)
(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) server-id)))
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)
;; if none running or if > 20 seconds since
;; server last used then start shutdown
;; This thread waits for the server to come alive
(debug:print-info 0 *default-log-port* "Starting the sync-back, keep alive thread in server")
(let* ((sdat #f)
(tmp-area (common:get-db-tmp-area))
(started-file (conc tmp-area "/.server-started"))
(server-start-time (current-seconds))
(server-info (let loop ((start-time (current-seconds))
(changed #t)
(last-sdat "not this"))
(begin ;; let ((sdat #f))
(thread-sleep! 0.01)
(debug:print-info 0 *default-log-port* "Waiting for server alive signature")
(mutex-lock! *heartbeat-mutex*)
(set! sdat *server-info*)
(mutex-unlock! *heartbeat-mutex*)
(if (and sdat
(not changed)
(> (- (current-seconds) start-time) 2))
(begin
(debug:print-info 0 *default-log-port* "Received server alive signature")
(common:save-pkt `((action . alive)
(T . server)
(pid . ,(current-process-id))
(ipaddr . ,(car sdat))
(port . ,(cadr sdat)))
*configdat* #t)
sdat)
(begin
(debug:print-info 0 *default-log-port* "Still waiting, last-sdat=" last-sdat)
(sleep 4)
(if (> (- (current-seconds) start-time) 120) ;; been waiting for two minutes
(begin
(debug:print-error 0 *default-log-port* "transport appears to have died, exiting server")
(common:save-pkt `((action . died)
(T . server)
(pid . ,(current-process-id))
(ipaddr . ,(car sdat))
(port . ,(cadr sdat))
(msg . "Transport died?"))
*configdat* #t)
(exit))
(loop start-time
(equal? sdat last-sdat)
sdat)))))))
(iface (car server-info))
(port (cadr server-info))
(last-access 0)
(server-timeout (server:expiration-timeout))
(server-going #f)
(server-log-file (args:get-arg "-log"))) ;; always set when we are a server
(handle-exceptions
exn
(debug:print 0 *default-log-port* "Failed to create " started-file ", exn=" exn)
(with-output-to-file started-file (lambda ()(print (current-process-id)))))
(let loop ((count 0)
(server-state 'available)
(bad-sync-count 0)
(start-time (current-milliseconds)))
;; Use this opportunity to sync the tmp db to megatest.db
(if (not server-going) ;; *dbstruct-db*
(begin
(debug:print 0 *default-log-port* "SERVER: dbprep")
(set! *dbstruct-db* (db:setup #t)) ;; run-id))
(set! server-going #t)
(debug:print 0 *default-log-port* "SERVER: running, megatest version: " (common:get-full-version)) ;; NOTE: the server is NOT yet marked as running in the log. We do that in the keep-running routine.
(thread-start! *watchdog*)))
;; when things go wrong we don't want to be doing the various queries too often
;; so we strive to run this stuff only every four seconds or so.
(let* ((sync-time (- (current-milliseconds) start-time))
(rem-time (quotient (- 4000 sync-time) 1000)))
(if (and (<= rem-time 4)
(> rem-time 0))
(thread-sleep! rem-time)))
(if (< count 1) ;; 3x3 = 9 secs aprox
(loop (+ count 1) 'running bad-sync-count (current-milliseconds)))
;; Check that iface and port have not changed (can happen if server port collides)
(mutex-lock! *heartbeat-mutex*)
(set! sdat *server-info*)
(mutex-unlock! *heartbeat-mutex*)
(if (not (equal? sdat (list iface port)))
(let ((new-iface (car sdat))
(new-port (cadr sdat)))
(debug:print-info 0 *default-log-port* "WARNING: interface changed, refreshing iface and port info")
(set! iface new-iface)
(set! port new-port)
(if (not *server-id*)
(set! *server-id* (server:mk-signature)))
(debug:print 0 *default-log-port* (current-seconds) (current-directory) (current-process-id) (argv))
(debug:print 0 *default-log-port* "SERVER STARTED: " iface ":" port " AT " (current-seconds) " server-id: " *server-id*)
(flush-output *default-log-port*)))
;; Transfer *db-last-access* to last-access to use in checking that we are still alive
(mutex-lock! *heartbeat-mutex*)
(set! last-access *db-last-access*)
(mutex-unlock! *heartbeat-mutex*)
(if (common:low-noise-print 120 (conc "server running on " iface ":" port))
(begin
(if (not *server-id*)
(set! *server-id* (server:mk-signature)))
(debug:print 0 *default-log-port* (current-seconds) (current-directory) (current-process-id) (argv))
(debug:print 0 *default-log-port* "SERVER STARTED: " iface ":" port " AT " (current-seconds) " server-id: " *server-id*)
(flush-output *default-log-port*)))
(if (common:low-noise-print 60 "dbstats")
(begin
(debug:print 0 *default-log-port* "Server stats:")
(db:print-current-query-stats)))
(let* ((hrs-since-start (/ (- (current-seconds) server-start-time) 3600)))
(cond
((and *server-run*
(> (+ last-access server-timeout)
(current-seconds)))
(if (common:low-noise-print 120 "server continuing")
(debug:print-info 0 *default-log-port* "Server continuing, seconds since last db access: " (- (current-seconds) last-access))
(let ((curr-time (current-seconds)))
(handle-exceptions
exn
(debug:print 0 *default-log-port* "ERROR: Failed to change timestamp on log file " server-log-file ". Are you out of space on that disk? exn=" exn)
(if (not *server-overloaded*)
(change-file-times server-log-file curr-time curr-time)))))
(loop 0 server-state bad-sync-count (current-milliseconds)))
(else
(debug:print-info 0 *default-log-port* "Server timed out. seconds since last db access: " (- (current-seconds) last-access))
(http-transport:server-shutdown port)))))))
(define (http-transport:server-shutdown port)
(begin
;;(BB> "http-transport:server-shutdown called")
(debug:print-info 0 *default-log-port* "Starting to shutdown the server. pid="(current-process-id))
;;
;; start_shutdown
;;
(set! *time-to-exit* #t) ;; tell on-exit to be fast as we've already cleaned up
(portlogger:open-run-close portlogger:set-port port "released")
(thread-sleep! 1)
;; (debug:print-info 0 *default-log-port* "Max cached queries was " *max-cache-size*)
;; (debug:print-info 0 *default-log-port* "Number of cached writes " *number-of-writes*)
;; (debug:print-info 0 *default-log-port* "Average cached write time "
;; (if (eq? *number-of-writes* 0)
;; "n/a (no writes)"
;; (/ *writes-total-delay*
;; *number-of-writes*))
;; " ms")
;; (debug:print-info 0 *default-log-port* "Number non-cached queries " *number-non-write-queries*)
;; (debug:print-info 0 *default-log-port* "Average non-cached time "
;; (if (eq? *number-non-write-queries* 0)
;; "n/a (no queries)"
;; (/ *total-non-write-delay*
;; *number-non-write-queries*))
;; " ms")
(db:print-current-query-stats)
(common:save-pkt `((action . exit)
(T . server)
(pid . ,(current-process-id)))
*configdat* #t)
(debug:print-info 0 *default-log-port* "Server shutdown complete. Exiting")
(exit)))
;; all routes though here end in exit ...
;;
;; start_server?
;;
(define (http-transport:launch)
;; check that a server start is in progress, pause or exit if so
(let* ((tmp-area (common:get-db-tmp-area))
(server-start (conc tmp-area "/.server-start"))
(server-started (conc tmp-area "/.server-started"))
(start-time (common:lazy-modification-time server-start))
(started-time (common:lazy-modification-time server-started))
(server-starting (< start-time started-time)) ;; if start-time is less than started-time then a server is still starting
(start-time-old (> (- (current-seconds) start-time) 5))
(cleanup-proc (lambda (msg)
(let* ((serv-fname (conc "server-" (current-process-id) "-" (get-host-name) ".log"))
(full-serv-fname (conc *toppath* "/logs/" serv-fname))
(new-serv-fname (conc *toppath* "/logs/" "defunct-" serv-fname)))
(debug:print 0 *default-log-port* msg)
(if (common:file-exists? full-serv-fname)
(system (conc "sleep 1;mv -f " full-serv-fname " " new-serv-fname))
(debug:print 0 *default-log-port* "INFO: cannot move " full-serv-fname " to " new-serv-fname))
(exit)))))
#;(if (and (not start-time-old) ;; last server start try was less than five seconds ago
(not server-starting))
(begin
(cleanup-proc "NOT starting server, there is either a recently started server or a server in process of starting")
(exit)))
;; lets not even bother to start if there are already three or more server files ready to go
#;(let* ((num-alive (server:get-num-alive (server:get-list *toppath*))))
(if (> num-alive 3)
(begin
(cleanup-proc (conc "ERROR: Aborting server start because there are already " num-alive " possible servers either running or starting up"))
(exit))))
(common:save-pkt `((action . start)
(T . server)
(pid . ,(current-process-id)))
*configdat* #t)
(let* ((th2 (make-thread (lambda ()
(debug:print-info 0 *default-log-port* "Server run thread started")
(http-transport:run
(if (args:get-arg "-server")
(args:get-arg "-server")
"-")
)) "Server run"))
(th3 (make-thread (lambda ()
(debug:print-info 0 *default-log-port* "Server monitor thread started")
(http-transport:keep-running)
"Keep running"))))
(thread-start! th2)
(thread-sleep! 0.25) ;; give the server time to settle before starting the keep-running monitor.
(thread-start! th3)
(set! *didsomething* #t)
(thread-join! th2)
(exit))))
;; (define (http-transport:server-signal-handler signum)
;; (signal-mask! signum)
;; (handle-exceptions
;; exn
;; (debug:print 0 *default-log-port* " ... exiting ...")
;; (let ((th1 (make-thread (lambda ()
;; (thread-sleep! 1))
;; "eat response"))
;; (th2 (make-thread (lambda ()
;; (debug:print-error 0 *default-log-port* "Received ^C, attempting clean exit. Please be patient and wait a few seconds before hitting ^C again.")
;; (thread-sleep! 3) ;; give the flush three seconds to do it's stuff
;; (debug:print 0 *default-log-port* " Done.")
;; (exit 4))
;; "exit on ^C timer")))
;; (thread-start! th2)
;; (thread-start! th1)
;; (thread-join! th2))))
|
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
|
84
85
86
87
88
89
90
91
|
;;======================================================================
;; S E R V E R U T I L I T I E S
;;======================================================================
;;======================================================================
;; C L I E N T S
;;======================================================================
|