Megatest

Diff
Login

Differences From Artifact [1e70529cd9]:

To Artifact [d3482e4682]:


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

(define (lock-queue:set-state dbdat test-id newstate #!key (remtries 10))
  (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200)
  (handle-exceptions
   exn
   (if (> remtries 0)
       (begin
	 (debug:print 0 "WARNING: exception on lock-queue:set-state. Trying again in 30 seconds.")
	 (debug:print 0 " message: " ((condition-property-accessor 'exn 'message) exn))
	 (thread-sleep! 30)
	 (lock-queue:set-state dbdat test-id newstate remtries: (- remtries 1)))
       (begin
	 (debug:print 0 "ERROR:  Failed to set lock state for test with id " test-id ", error: " ((condition-property-accessor 'exn 'message) exn) ", giving up.")
	 #f))
   (sqlite3:execute (lock-queue:db-dat-get-db dbdat) "UPDATE queue SET state=? WHERE test_id=?;"
		    newstate
		    test-id)))

(define (lock-queue:any-younger? dbdat mystart test-id #!key (remtries 10))
  ;; no need to wait on journal on read only queries
  ;; (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200)
  (handle-exceptions
   exn
   (if (> remtries 0)
       (begin
	 (debug:print 0 "WARNING: exception on lock-queue:any-younger. Removing lockdb and trying again in 5 seconds.")
	 (debug:print 0 " message: " ((condition-property-accessor 'exn 'message) exn))
	 (thread-sleep! 5)
         (lock-queue:delete-lock-db dbdat)
	 (lock-queue:any-younger? dbdat mystart test-id remtries: (- remtries 1)))
       (begin
	 (debug:print 0 "ERROR:  Failed to find younger locks for test with id " test-id ", error: " ((condition-property-accessor 'exn 'message) exn) ", giving up.")
	 #f))
   (let ((res #f))
     (sqlite3:for-each-row
      (lambda (tid)
	;; Actually this should not be needed as mystart cannot be simultaneously less than and test-id same as 
	(if (not (equal? tid test-id)) 
	    (set! res tid)))







|
|



|












|
|




|







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

(define (lock-queue:set-state dbdat test-id newstate #!key (remtries 10))
  (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200)
  (handle-exceptions
   exn
   (if (> remtries 0)
       (begin
	 (debug:print 0 #f "WARNING: exception on lock-queue:set-state. Trying again in 30 seconds.")
	 (debug:print 0 #f " message: " ((condition-property-accessor 'exn 'message) exn))
	 (thread-sleep! 30)
	 (lock-queue:set-state dbdat test-id newstate remtries: (- remtries 1)))
       (begin
	 (debug:print 0 #f "ERROR:  Failed to set lock state for test with id " test-id ", error: " ((condition-property-accessor 'exn 'message) exn) ", giving up.")
	 #f))
   (sqlite3:execute (lock-queue:db-dat-get-db dbdat) "UPDATE queue SET state=? WHERE test_id=?;"
		    newstate
		    test-id)))

(define (lock-queue:any-younger? dbdat mystart test-id #!key (remtries 10))
  ;; no need to wait on journal on read only queries
  ;; (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200)
  (handle-exceptions
   exn
   (if (> remtries 0)
       (begin
	 (debug:print 0 #f "WARNING: exception on lock-queue:any-younger. Removing lockdb and trying again in 5 seconds.")
	 (debug:print 0 #f " message: " ((condition-property-accessor 'exn 'message) exn))
	 (thread-sleep! 5)
         (lock-queue:delete-lock-db dbdat)
	 (lock-queue:any-younger? dbdat mystart test-id remtries: (- remtries 1)))
       (begin
	 (debug:print 0 #f "ERROR:  Failed to find younger locks for test with id " test-id ", error: " ((condition-property-accessor 'exn 'message) exn) ", giving up.")
	 #f))
   (let ((res #f))
     (sqlite3:for-each-row
      (lambda (tid)
	;; Actually this should not be needed as mystart cannot be simultaneously less than and test-id same as 
	(if (not (equal? tid test-id)) 
	    (set! res tid)))
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
	 (db        (lock-queue:db-dat-get-db dbdat))
	 (lckqry    (sqlite3:prepare db "SELECT test_id,run_lock FROM runlocks WHERE run_lock='locked';"))
	 (mklckqry  (sqlite3:prepare db "INSERT INTO runlocks (test_id,run_lock) VALUES (?,'locked');")))
    (let ((result 
	   (handle-exceptions
	    exn
	    (begin
	      (debug:print 0 "WARNING: failed to get queue lock. Removing lock db and returning fail") ;; Will try again in a few seconds")
	      (debug:print 0 " message: " ((condition-property-accessor 'exn 'message) exn))
	      (thread-sleep! 10)
	      ;; (if (> count 0)	
	      ;;  #f ;; (lock-queue:get-lock dbdat test-id count: (- count 1)) - give up on retries 
	      ;; (begin ;; never recovered, remote the lock file and return #f, no lock obtained
	      (lock-queue:delete-lock-db dbdat)
	      #f)
	    (sqlite3:with-transaction







|
|







117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
	 (db        (lock-queue:db-dat-get-db dbdat))
	 (lckqry    (sqlite3:prepare db "SELECT test_id,run_lock FROM runlocks WHERE run_lock='locked';"))
	 (mklckqry  (sqlite3:prepare db "INSERT INTO runlocks (test_id,run_lock) VALUES (?,'locked');")))
    (let ((result 
	   (handle-exceptions
	    exn
	    (begin
	      (debug:print 0 #f "WARNING: failed to get queue lock. Removing lock db and returning fail") ;; Will try again in a few seconds")
	      (debug:print 0 #f " message: " ((condition-property-accessor 'exn 'message) exn))
	      (thread-sleep! 10)
	      ;; (if (> count 0)	
	      ;;  #f ;; (lock-queue:get-lock dbdat test-id count: (- count 1)) - give up on retries 
	      ;; (begin ;; never recovered, remote the lock file and return #f, no lock obtained
	      (lock-queue:delete-lock-db dbdat)
	      #f)
	    (sqlite3:with-transaction
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

(define (lock-queue:release-lock fname test-id #!key (count 10))
  (let* ((dbdat (lock-queue:open-db fname)))
    (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200 "lock-queue:release-lock; waiting on journal")
    (handle-exceptions
     exn
     (begin
       (debug:print 0 "WARNING: Failed to release queue lock. Will try again in few seconds")
       (debug:print 0 " message: " ((condition-property-accessor 'exn 'message) exn))
       (thread-sleep! (/ count 10))
       (if (> count 0)
	   (begin
	     (sqlite3:finalize! (lock-queue:db-dat-get-db dbdat))
	     (lock-queue:release-lock fname test-id count: (- count 1)))
	   (let ((journal (conc fname "-journal")))
	     ;; If we've tried ten times and failed there is a serious problem
	     ;; try to remove the lock db and allow it to be recreated
	     (handle-exceptions
	      exn
	      #f
	      (if (file-exists? journal)(delete-file journal))
	      (if (file-exists? fname)  (delete-file fname))
	      #f))))
     (sqlite3:execute (lock-queue:db-dat-get-db dbdat) "DELETE FROM runlocks WHERE test_id=?;" test-id)
     (sqlite3:finalize! (lock-queue:db-dat-get-db dbdat)))))

(define (lock-queue:steal-lock dbdat test-id #!key (count 10))
  (debug:print-info 0 "Attempting to steal lock at " (lock-queue:db-dat-get-path dbdat))
  (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200 "lock-queue:steal-lock; waiting on journal")
  (handle-exceptions
   exn
   (begin
     (tadebug:print 0 "WARNING: Failed to steal queue lock. Will try again in few seconds")
     (debug:print 0 " message: " ((condition-property-accessor 'exn 'message) exn))
     (thread-sleep! 10)
     (if (> count 0)
	 (lock-queue:steal-lock dbdat test-id count: (- count 1))
	 #f))
   (sqlite3:execute (lock-queue:db-dat-get-db dbdat) "DELETE FROM runlocks WHERE run_lock='locked';"))
  (lock-queue:get-lock dbdat test-it))

;; returns #f if ok to skip the task
;; returns #t if ok to proceed with task
;; otherwise waits
;;
(define (lock-queue:wait-turn fname test-id #!key (count 10)(waiting-msg #f))
  (let* ((dbdat   (lock-queue:open-db fname))
	 (mystart (current-seconds))
	 (db      (lock-queue:db-dat-get-db dbdat)))
    ;; (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200 waiting-msg: "lock-queue:wait-turn; waiting on journal file")
    (handle-exceptions
     exn
     (begin
       (debug:print 0 "WARNING: Failed to find out if it is ok to skip the wait queue. Will try again in few seconds")
       (debug:print 0 " message: " ((condition-property-accessor 'exn 'message) exn))
       (print-call-chain (current-error-port))
       (thread-sleep! 10)
       (if (> count 0)
	   (begin
	     (sqlite3:finalize! db)
	     (lock-queue:wait-turn fname test-id count: (- count 1)))
	   (begin
	     (debug:print 0 "Giving up calls to lock-queue:wait-turn for test-id " test-id " at path " fname ", printing call chain")
	     (print-call-chain (current-error-port))
	     #f)))
     ;; wait 10 seconds and then check to see if someone is already updating the html
     (thread-sleep! 10)
     (if (not (lock-queue:any-younger? dbdat mystart test-id)) ;; no processing in flight, must try to start processing
	 (begin
	   (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200 waiting-msg: "lock-queue:wait-turn; waiting on journal file")







|
|


















|




|
|



















|
|







|







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

(define (lock-queue:release-lock fname test-id #!key (count 10))
  (let* ((dbdat (lock-queue:open-db fname)))
    (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200 "lock-queue:release-lock; waiting on journal")
    (handle-exceptions
     exn
     (begin
       (debug:print 0 #f "WARNING: Failed to release queue lock. Will try again in few seconds")
       (debug:print 0 #f " message: " ((condition-property-accessor 'exn 'message) exn))
       (thread-sleep! (/ count 10))
       (if (> count 0)
	   (begin
	     (sqlite3:finalize! (lock-queue:db-dat-get-db dbdat))
	     (lock-queue:release-lock fname test-id count: (- count 1)))
	   (let ((journal (conc fname "-journal")))
	     ;; If we've tried ten times and failed there is a serious problem
	     ;; try to remove the lock db and allow it to be recreated
	     (handle-exceptions
	      exn
	      #f
	      (if (file-exists? journal)(delete-file journal))
	      (if (file-exists? fname)  (delete-file fname))
	      #f))))
     (sqlite3:execute (lock-queue:db-dat-get-db dbdat) "DELETE FROM runlocks WHERE test_id=?;" test-id)
     (sqlite3:finalize! (lock-queue:db-dat-get-db dbdat)))))

(define (lock-queue:steal-lock dbdat test-id #!key (count 10))
  (debug:print-info 0 #f "Attempting to steal lock at " (lock-queue:db-dat-get-path dbdat))
  (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200 "lock-queue:steal-lock; waiting on journal")
  (handle-exceptions
   exn
   (begin
     (debug:print 0 #f "WARNING: Failed to steal queue lock. Will try again in few seconds")
     (debug:print 0 #f " message: " ((condition-property-accessor 'exn 'message) exn))
     (thread-sleep! 10)
     (if (> count 0)
	 (lock-queue:steal-lock dbdat test-id count: (- count 1))
	 #f))
   (sqlite3:execute (lock-queue:db-dat-get-db dbdat) "DELETE FROM runlocks WHERE run_lock='locked';"))
  (lock-queue:get-lock dbdat test-it))

;; returns #f if ok to skip the task
;; returns #t if ok to proceed with task
;; otherwise waits
;;
(define (lock-queue:wait-turn fname test-id #!key (count 10)(waiting-msg #f))
  (let* ((dbdat   (lock-queue:open-db fname))
	 (mystart (current-seconds))
	 (db      (lock-queue:db-dat-get-db dbdat)))
    ;; (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200 waiting-msg: "lock-queue:wait-turn; waiting on journal file")
    (handle-exceptions
     exn
     (begin
       (debug:print 0 #f "WARNING: Failed to find out if it is ok to skip the wait queue. Will try again in few seconds")
       (debug:print 0 #f " message: " ((condition-property-accessor 'exn 'message) exn))
       (print-call-chain (current-error-port))
       (thread-sleep! 10)
       (if (> count 0)
	   (begin
	     (sqlite3:finalize! db)
	     (lock-queue:wait-turn fname test-id count: (- count 1)))
	   (begin
	     (debug:print 0 #f "Giving up calls to lock-queue:wait-turn for test-id " test-id " at path " fname ", printing call chain")
	     (print-call-chain (current-error-port))
	     #f)))
     ;; wait 10 seconds and then check to see if someone is already updating the html
     (thread-sleep! 10)
     (if (not (lock-queue:any-younger? dbdat mystart test-id)) ;; no processing in flight, must try to start processing
	 (begin
	   (tasks:wait-on-journal (lock-queue:db-dat-get-path dbdat) 1200 waiting-msg: "lock-queue:wait-turn; waiting on journal file")