34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
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
|
;; grep define ../rmt.scm | grep rmt: |perl -pi -e 's/\(define\s+\((\S+)\W.*$/\1/'|sort -u
;;======================================================================
;; S U P P O R T F U N C T I O N S
;;======================================================================
;;
(define (rmt:write-frequency-over-limit? cmd run-id)
(and (not (member cmd api:read-only-queries))
(let* ((tmprec (hash-table-ref/default *write-frequency* run-id #f))
(record (if tmprec tmprec
(let ((v (vector (current-seconds) 0)))
(hash-table-set! *write-frequency* run-id v)
v)))
(count (+ 1 (vector-ref record 1)))
(start (vector-ref record 0))
(queries-per-second (/ (* count 1.0)
(max (- (current-seconds) start) 1))))
(vector-set! record 1 count)
(if (and (> count 10)
(> queries-per-second 10))
(begin
(debug:print-info 1 *default-log-port* "db write rate too high, starting a server, count=" count " start=" start " run-id=" run-id " queries-per-second=" queries-per-second)
#t)
#f))))
(let ((cinfo (hash-table-ref/default *runremote* run-id #f)))
(if cinfo
cinfo
;; NB// can cache the answer for server running for 10 seconds ...
;; ;; (and (not (rmt:write-frequency-over-limit? cmd run-id))
(let ((expire-time (- (current-seconds) (server:get-timeout) 10))) ;; don't forget the 10 second margin
(for-each
(lambda (run-id)
(let ((connection (hash-table-ref/default *runremote* run-id #f)))
(if (and (vector? connection)
(< (http-transport:server-dat-get-last-access connection) expire-time))
(begin
(debug:print-info 0 *default-log-port* "Discarding connection to server for run-id " run-id ", too long between accesses")
;; bb- disabling nanomsg
;; SHOULD CLOSE THE CONNECTION HERE
;; (case *transport-type*
;; ((nmsg)(nn-close (http-transport:server-dat-get-socket
;; (hash-table-ref *runremote* run-id)))))
(hash-table-delete! *runremote* run-id)))))
(hash-table-keys *runremote*)))
;; (mutex-unlock! *db-multi-sync-mutex*)
;; (mutex-lock! *send-receive-mutex*)
(let* ((run-id (if rid rid 0))
(connection-info (rmt:get-connection-info run-id)))
;; the nmsg method does the encoding under the hood (the http method should be changed to do this also)
(if connection-info
;; use the server if have connection info
(let* ((dat (case *transport-type*
((http)(condition-case
(http-transport:client-api-send-receive run-id connection-info cmd params)
((commfail)(vector #f "communications fail"))
((exn)(vector #f "other fail"))))
;; ((nmsg)(condition-case
;; (nmsg-transport:client-api-send-receive run-id connection-info cmd params)
;; ((timeout)(vector #f "timeout talking to server"))))
(else (exit))))
(success (if (vector? dat) (vector-ref dat 0) #f))
(res (if (vector? dat) (vector-ref dat 1) #f)))
(if (vector? connection-info)(http-transport:server-dat-update-last-access connection-info))
(if success
(begin
;; (mutex-unlock! *send-receive-mutex*)
(case *transport-type*
((http) res) ;; (db:string->obj res))
;; ((nmsg) res)
)) ;; (vector-ref res 1)))
(begin ;; let ((new-connection-info (client:setup run-id)))
(debug:print 0 *default-log-port* "WARNING: Communication failed, trying call to rmt:send-receive again.")
;; (case *transport-type*
;; ((nmsg)(nn-close (http-transport:server-dat-get-socket connection-info))))
(hash-table-delete! *runremote* run-id) ;; don't keep using the same connection
;; NOTE: killing server causes this process to block forever. No idea why. Dec 2.
;; (if (eq? (modulo attemptnum 5) 0)
;; (tasks:kill-server-run-id run-id tag: "api-send-receive-failed"))
;; (mutex-unlock! *send-receive-mutex*) ;; close the mutex here to allow other threads access to communications
(tasks:start-and-wait-for-server (tasks:open-db) run-id 15)
;; (nmsg-transport:client-api-send-receive run-id connection-info cmd param remtries: (- remtries 1))))))
;; no longer killing the server in http-transport:client-api-send-receive
;; may kill it here but what are the criteria?
;; start with three calls then kill server
;; (if (eq? attemptnum 3)(tasks:kill-server-run-id run-id))
;; (thread-sleep! 2)
(rmt:send-receive cmd run-id params attemptnum: (+ attemptnum 1)))))
;; no connection info? try to start a server, or access locally if no
;; server and the query is read-only
;;
;; Note: The tasks db was checked for a server in starting mode in the rmt:get-connection-info call
;;
(if (and (< attemptnum 15)
(member cmd api:write-queries))
(let ((faststart (configf:lookup *configdat* "server" "faststart")))
(hash-table-delete! *runremote* run-id)
;; (mutex-unlock! *send-receive-mutex*)
(if (and faststart (equal? faststart "no"))
(begin
(tasks:start-and-wait-for-server (db:delay-if-busy (tasks:open-db)) run-id 10)
(thread-sleep! (random 5)) ;; give some time to settle and minimize collison?
(rmt:send-receive cmd rid params attemptnum: (+ attemptnum 1)))
(let ((start-time (current-milliseconds))
(max-query (string->number (or (configf:lookup *configdat* "server" "server-query-threshold")
"300")))
(newres (rmt:open-qry-close-locally cmd run-id params)))
(let ((delta (- (current-milliseconds) start-time)))
(if (> delta max-query)
(begin
(debug:print-info 0 *default-log-port* "Starting server as query time " delta " is over the limit of " max-query)
(server:kind-run run-id)))
;; return the result!
newres)
)))
(begin
;; (debug:print-error 0 *default-log-port* "Communication failed!")
;; (mutex-unlock! *send-receive-mutex*)
;; (exit)
(rmt:open-qry-close-locally cmd run-id params)
)))))
(define (rmt:update-db-stats run-id rawcmd params duration)
(mutex-lock! *db-stats-mutex*)
(handle-exceptions
exn
(begin
(debug:print 0 *default-log-port* "WARNING: stats collection failed in update-db-stats")
|
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
|
<
<
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
>
>
>
|
>
>
>
|
<
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
<
<
<
<
<
<
|
>
>
>
>
>
>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
>
|
|
|
|
|
|
|
|
|
>
|
|
|
|
|
|
|
|
|
|
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
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
|
;; grep define ../rmt.scm | grep rmt: |perl -pi -e 's/\(define\s+\((\S+)\W.*$/\1/'|sort -u
;;======================================================================
;; S U P P O R T F U N C T I O N S
;;======================================================================
;; if a server is either running or in the process of starting call client:setup
;; else return #f to let the calling proc know that there is no server available
;;
(define (rmt:get-connection-info run-id)
(let ((cinfo *runremote*)) ;; (hash-table-ref/default *runremote* run-id #f)))
(if cinfo
cinfo
(if (tasks:server-running-or-starting? (db:delay-if-busy (tasks:open-db)) run-id)
(client:setup run-id)
#f))))
(define *send-receive-mutex* (make-mutex)) ;; should have separate mutex per run-id
;; RA => e.g. usage (rmt:send-receive 'get-var #f (list varname))
;;
(define (rmt:send-receive cmd rid params #!key (attemptnum 1)) ;; start attemptnum at 1 so the modulo below works as expected
;; clean out old connections
;; (mutex-lock! *db-multi-sync-mutex*)
;; 1. check if server is started IFF cmd is a write OR if we are not on the homehost, store in *runremote*
;; 2. check the age of the connections. refresh the connection if it is older than timeout-20 seconds.
;; 3. do the query, if on homehost use local access
(if (and #f ;; FORCE NO GO FOR RIGHT NOW
(not *runremote*) ;; we trust *runremote* to reflect that a server was found previously
(not (member cmd api:read-only-queries))) ;; we don't trust so much the list of write queries
(let ((serverconn (server:check-if-running *toppath*)))
(if serverconn
(set! *runremote* serverconn) ;; the string can be consumed by the client setup if needed
(server:kind-run *toppath*))))
(rmt:open-qry-close-locally cmd (if rid rid 0) params))
;; (let ((expire-time (- (current-seconds) (server:get-timeout) 10))) ;; don't forget the 10 second margin
;; (for-each
;; (lambda (run-id)
;; (let ((connection (hash-table-ref/default *runremote* run-id #f)))
;; (if (and (vector? connection)
;; (< (http-transport:server-dat-get-last-access connection) expire-time))
;; (begin
;; (debug:print-info 0 *default-log-port* "Discarding connection to server for run-id " run-id ", too long between accesses")
;; ;; bb- disabling nanomsg
;; ;; SHOULD CLOSE THE CONNECTION HERE
;; ;; (case *transport-type*
;; ;; ((nmsg)(nn-close (http-transport:server-dat-get-socket
;; ;; (hash-table-ref *runremote* run-id)))))
;; (hash-table-delete! *runremote* run-id)))))
;; (hash-table-keys *runremote*)))
;; ;; (mutex-unlock! *db-multi-sync-mutex*)
;; ;; (mutex-lock! *send-receive-mutex*)
;; (let* ((run-id (if rid rid 0))
;; (home-host (common:get-homehost))
;; (connection-info (if (cdr home-host) ;; we are on the home-host
;; #f
;; (rmt:get-connection-info run-id))))
;; (cond
;; (home-host (rmt:open-qry-close-locally cmd run-id params))
;; (connection-info
;; ;; the nmsg method does the encoding under the hood (the http method should be changed to do this also)
;; ;; use the server if have connection info
;; (let* ((dat (case *transport-type*
;; ((http)(condition-case
;; (http-transport:client-api-send-receive run-id connection-info cmd params)
;; ((commfail)(vector #f "communications fail"))
;; ((exn)(vector #f "other fail"))))
;; ;; ((nmsg)(condition-case
;; ;; (nmsg-transport:client-api-send-receive run-id connection-info cmd params)
;; ;; ((timeout)(vector #f "timeout talking to server"))))
;; (else (exit))))
;; (success (if (vector? dat) (vector-ref dat 0) #f))
;; (res (if (vector? dat) (vector-ref dat 1) #f)))
;; (if (vector? connection-info)(http-transport:server-dat-update-last-access connection-info))
;; (if success
;; (begin
;; ;; (mutex-unlock! *send-receive-mutex*)
;; (case *transport-type*
;; ((http) res) ;; (db:string->obj res))
;; ;; ((nmsg) res)
;; )) ;; (vector-ref res 1)))
;; (begin ;; let ((new-connection-info (client:setup run-id)))
;; (debug:print 0 *default-log-port* "WARNING: Communication failed, trying call to rmt:send-receive again.")
;; ;; (case *transport-type*
;; ;; ((nmsg)(nn-close (http-transport:server-dat-get-socket connection-info))))
;; (hash-table-delete! *runremote* run-id) ;; don't keep using the same connection
;; ;; NOTE: killing server causes this process to block forever. No idea why. Dec 2.
;; ;; (if (eq? (modulo attemptnum 5) 0)
;; ;; (tasks:kill-server-run-id run-id tag: "api-send-receive-failed"))
;; ;; (mutex-unlock! *send-receive-mutex*) ;; close the mutex here to allow other threads access to communications
;; (tasks:start-and-wait-for-server (tasks:open-db) run-id 15)
;; ;; (nmsg-transport:client-api-send-receive run-id connection-info cmd param remtries: (- remtries 1))))))
;;
;; ;; no longer killing the server in http-transport:client-api-send-receive
;; ;; may kill it here but what are the criteria?
;; ;; start with three calls then kill server
;; ;; (if (eq? attemptnum 3)(tasks:kill-server-run-id run-id))
;; ;; (thread-sleep! 2)
;; (rmt:send-receive cmd run-id params attemptnum: (+ attemptnum 1))))))
;; (else
;; ;; no connection info? try to start a server, or access locally if no
;; ;; server and the query is read-only
;; ;;
;; ;; Note: The tasks db was checked for a server in starting mode in the rmt:get-connection-info call
;; ;;
;; (if (and (< attemptnum 15)
;; (member cmd api:write-queries))
;; (let ((homehost (common:get-homehost))) ;; faststart (configf:lookup *configdat* "server" "faststart")))
;; (hash-table-delete! *runremote* run-id)
;; ;; (mutex-unlock! *send-receive-mutex*)
;; (if (not (cdr homehost)) ;; we always require a server if not on homehost ;; (and faststart (equal? faststart "no"))
;; (begin
;; (tasks:start-and-wait-for-server (db:delay-if-busy (tasks:open-db)) run-id 10)
;; (thread-sleep! (random 5)) ;; give some time to settle and minimize collison?
;; (rmt:send-receive cmd rid params attemptnum: (+ attemptnum 1)))
;; ;; NB - probably can remove the query time stuff but need to discuss it ....
;; (let ((start-time (current-milliseconds))
;; (max-query (string->number (or (configf:lookup *configdat* "server" "server-query-threshold")
;; "300")))
;; (newres (rmt:open-qry-close-locally cmd run-id params)))
;; (let ((delta (- (current-milliseconds) start-time)))
;; (if (> delta max-query)
;; (begin
;; (debug:print-info 0 *default-log-port* "WARNING: long query times, you may have an overloaded homehost.") ;; Starting server as query time " delta " is over the limit of " max-query)
;; ;; (server:kind-run run-id)))
;; ))
;; ;; return the result!
;; newres)
;; )))
;; (begin
;; ;; (debug:print-error 0 *default-log-port* "Communication failed!")
;; ;; (mutex-unlock! *send-receive-mutex*)
;; ;; (exit)
;; (rmt:open-qry-close-locally cmd run-id params)
;; ))))))
(define (rmt:update-db-stats run-id rawcmd params duration)
(mutex-lock! *db-stats-mutex*)
(handle-exceptions
exn
(begin
(debug:print 0 *default-log-port* "WARNING: stats collection failed in update-db-stats")
|
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
|
(cons newmax-cmd currmax)
(cons 'none 0))
(loop (car tal)(cdr tal) newmax-cmd currmax)))))))
(mutex-unlock! *db-stats-mutex*)
res))
(define (rmt:open-qry-close-locally cmd run-id params #!key (remretries 5))
(let* ((dbstruct-local (db:open-local-db-handle))
(db-file-path (db:dbfile-path 0))
;; (read-only (not (file-read-access? db-file-path)))
(start (current-milliseconds))
(resdat (api:execute-requests dbstruct-local (vector (symbol->string cmd) params)))
(success (vector-ref resdat 0))
(res (vector-ref resdat 1))
(duration (- (current-milliseconds) start)))
(if (not success)
(if (> remretries 0)
(begin
(debug:print-error 0 *default-log-port* "local query failed. Trying again.")
(thread-sleep! (/ (random 5000) 1000)) ;; some random delay
(rmt:open-qry-close-locally cmd run-id params remretries: (- remretries 1)))
(begin
(debug:print-error 0 *default-log-port* "too many retries in rmt:open-qry-close-locally, giving up")
#f))
(begin
;; (rmt:update-db-stats run-id cmd params duration)
;; mark this run as dirty if this was a write
(if (not (member cmd api:read-only-queries))
(let ((start-time (current-seconds)))
(mutex-lock! *db-multi-sync-mutex*)
;; (if (not (hash-table-ref/default *db-local-sync* run-id #f))
;; just set it every time. Is a write more expensive than a read and does it matter?
(mutex-unlock! *db-multi-sync-mutex*)))
res))))
|
|
|
>
|
>
|
>
>
>
|
|
<
<
|
|
|
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
|
(cons newmax-cmd currmax)
(cons 'none 0))
(loop (car tal)(cdr tal) newmax-cmd currmax)))))))
(mutex-unlock! *db-stats-mutex*)
res))
(define (rmt:open-qry-close-locally cmd run-id params #!key (remretries 5))
(let* ((qry-is-write (not (member cmd api:read-only-queries)))
(db-file-path (db:dbfile-path)) ;; 0))
(dbstruct-local (db:setup)) ;; make-dbr:dbstruct path: dbdir local: #t)))
(read-only (not (file-write-access? db-file-path)))
(start (current-milliseconds))
(resdat (if (not (and read-only qry-is-write))
(api:execute-requests dbstruct-local (vector (symbol->string cmd) params))
(vector #t '())))
(success (vector-ref resdat 0))
(res (vector-ref resdat 1))
(duration (- (current-milliseconds) start)))
(if (and read-only qry-is-write)
(debug:print 0 *default-log-port* "ERROR: attempt to write to read-only database ignored. cmd=" cmd))
(if (not success)
(if (> remretries 0)
(begin
(debug:print-error 0 *default-log-port* "local query failed. Trying again.")
(thread-sleep! (/ (random 5000) 1000)) ;; some random delay
(rmt:open-qry-close-locally cmd run-id params remretries: (- remretries 1)))
(begin
(debug:print-error 0 *default-log-port* "too many retries in rmt:open-qry-close-locally, giving up")
#f))
(begin
;; (rmt:update-db-stats run-id cmd params duration)
;; mark this run as dirty if this was a write, the watchdog is responsible for syncing it
(if qry-is-write
(let ((start-time (current-seconds)))
(mutex-lock! *db-multi-sync-mutex*)
(hash-table-set! *db-local-sync* (or run-id 0) start-time) ;; the oldest "write"
(mutex-unlock! *db-multi-sync-mutex*)))))
res))
(define (rmt:send-receive-no-auto-client-setup connection-info cmd run-id params)
(let* ((run-id (if run-id run-id 0))
;; (jparams (db:obj->string params)) ;; (rmt:dat->json-str params))
(res (handle-exceptions
exn
#f
|
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
|
(define (rmt:login run-id)
(rmt:send-receive 'login run-id (list *toppath* megatest-version run-id *my-client-signature*)))
;; This login does no retries under the hood - it acts a bit like a ping.
;; Deprecated for nmsg-transport.
;;
(define (rmt:login-no-auto-client-setup connection-info run-id)
(case *transport-type*
((http)(rmt:send-receive-no-auto-client-setup connection-info 'login run-id (list *toppath* megatest-version run-id *my-client-signature*)))
;;((nmsg)(nmsg-transport:client-api-send-receive run-id connection-info 'login (list *toppath* megatest-version run-id *my-client-signature*)))
))
;; hand off a call to one of the db:queries statements
;; added run-id to make looking up the correct db possible
;;
(define (rmt:general-call stmtname run-id . params)
|
|
|
|
|
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
|
(define (rmt:login run-id)
(rmt:send-receive 'login run-id (list *toppath* megatest-version run-id *my-client-signature*)))
;; This login does no retries under the hood - it acts a bit like a ping.
;; Deprecated for nmsg-transport.
;;
(define (rmt:login-no-auto-client-setup connection-info)
(case *transport-type* ;; run-id of 0 is just a placeholder
((http)(rmt:send-receive-no-auto-client-setup connection-info 'login 0 (list *toppath* megatest-version *my-client-signature*)))
;;((nmsg)(nmsg-transport:client-api-send-receive run-id connection-info 'login (list *toppath* megatest-version run-id *my-client-signature*)))
))
;; hand off a call to one of the db:queries statements
;; added run-id to make looking up the correct db possible
;;
(define (rmt:general-call stmtname run-id . params)
|
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
|
(rmt:send-receive 'get-count-tests-running-for-testname run-id (list run-id testname)))
(define (rmt:get-count-tests-running-in-jobgroup run-id jobgroup)
(rmt:send-receive 'get-count-tests-running-in-jobgroup run-id (list run-id jobgroup)))
;; state and status are extra hints not usually used in the calculation
;;
(define (rmt:roll-up-pass-fail-counts run-id test-name item-path state status)
(rmt:send-receive 'roll-up-pass-fail-counts run-id (list run-id test-name item-path state status)))
(define (rmt:update-pass-fail-counts run-id test-name)
(rmt:general-call 'update-pass-fail-counts run-id test-name test-name test-name))
(define (rmt:top-test-set-per-pf-counts run-id test-name)
(rmt:send-receive 'top-test-set-per-pf-counts run-id (list run-id test-name)))
|
|
|
|
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
|
(rmt:send-receive 'get-count-tests-running-for-testname run-id (list run-id testname)))
(define (rmt:get-count-tests-running-in-jobgroup run-id jobgroup)
(rmt:send-receive 'get-count-tests-running-in-jobgroup run-id (list run-id jobgroup)))
;; state and status are extra hints not usually used in the calculation
;;
(define (rmt:roll-up-pass-fail-counts run-id test-name item-path state status comment)
(rmt:send-receive 'roll-up-pass-fail-counts run-id (list run-id test-name item-path state status comment)))
(define (rmt:update-pass-fail-counts run-id test-name)
(rmt:general-call 'update-pass-fail-counts run-id test-name test-name test-name))
(define (rmt:top-test-set-per-pf-counts run-id test-name)
(rmt:send-receive 'top-test-set-per-pf-counts run-id (list run-id test-name)))
|