| ︙ | | | ︙ | |
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
|
(launch:einf-exit-code-set! exit-info exit-code) ;; (vector-set! exit-info 2 exit-code)
(mutex-unlock! m)
(if (eq? pid-val 0)
(begin
(thread-sleep! 2)
(processloop (+ i 1))))
)))))
(debug:print-info 0 #f "step " stepname " completed with exit code " (launch:einf-exit-code exit-info)) ;; (vector-ref exit-info 2))
;; now run logpro if needed
(if logpro-used
(let ((pid (process-run (conc "logpro " logpro-file " " (conc stepname ".html") " < " stepname ".log"))))
(let processloop ((i 0))
(let-values (((pid-val exit-status exit-code)(process-wait pid #t)))
(mutex-lock! m)
;; (make-launch:einf pid: pid exit-status: exit-status exit-code: exit-code)
(launch:einf-pid-set! exit-info pid) ;; (vector-set! exit-info 0 pid)
(launch:einf-exit-status-set! exit-info exit-status) ;; (vector-set! exit-info 1 exit-status)
(launch:einf-exit-code-set! exit-info exit-code) ;; (vector-set! exit-info 2 exit-code)
(mutex-unlock! m)
(if (eq? pid-val 0)
(begin
(thread-sleep! 2)
(processloop (+ i 1)))))
(debug:print-info 0 #f "logpro for step " stepname " exited with code " (launch:einf-exit-code exit-info))))) ;; (vector-ref exit-info 2)))))
(let ((exinfo (launch:einf-exit-code exit-info)) ;; (vector-ref exit-info 2))
(logfna (if logpro-used (conc stepname ".html") ""))
(comment #f))
(if logpro-used
(let ((datfile (conc stepname ".dat")))
;; load the .dat file into the test_data table if it exists
|
|
|
|
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
|
(launch:einf-exit-code-set! exit-info exit-code) ;; (vector-set! exit-info 2 exit-code)
(mutex-unlock! m)
(if (eq? pid-val 0)
(begin
(thread-sleep! 2)
(processloop (+ i 1))))
)))))
(debug:print-info 0 *default-log-port* "step " stepname " completed with exit code " (launch:einf-exit-code exit-info)) ;; (vector-ref exit-info 2))
;; now run logpro if needed
(if logpro-used
(let ((pid (process-run (conc "logpro " logpro-file " " (conc stepname ".html") " < " stepname ".log"))))
(let processloop ((i 0))
(let-values (((pid-val exit-status exit-code)(process-wait pid #t)))
(mutex-lock! m)
;; (make-launch:einf pid: pid exit-status: exit-status exit-code: exit-code)
(launch:einf-pid-set! exit-info pid) ;; (vector-set! exit-info 0 pid)
(launch:einf-exit-status-set! exit-info exit-status) ;; (vector-set! exit-info 1 exit-status)
(launch:einf-exit-code-set! exit-info exit-code) ;; (vector-set! exit-info 2 exit-code)
(mutex-unlock! m)
(if (eq? pid-val 0)
(begin
(thread-sleep! 2)
(processloop (+ i 1)))))
(debug:print-info 0 *default-log-port* "logpro for step " stepname " exited with code " (launch:einf-exit-code exit-info))))) ;; (vector-ref exit-info 2)))))
(let ((exinfo (launch:einf-exit-code exit-info)) ;; (vector-ref exit-info 2))
(logfna (if logpro-used (conc stepname ".html") ""))
(comment #f))
(if logpro-used
(let ((datfile (conc stepname ".dat")))
;; load the .dat file into the test_data table if it exists
|
| ︙ | | | ︙ | |
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
|
df
#f))))
(set! kill-job? (or (test-get-kill-request run-id test-id) ;; run-id test-name itemdat))
(and runtlim (let* ((run-seconds (- (current-seconds) start-seconds))
(time-exceeded (> run-seconds runtlim)))
(if time-exceeded
(begin
(debug:print-info 0 #f "KILLING TEST DUE TO TIME LIMIT EXCEEDED! Runtime=" run-seconds " seconds, limit=" runtlim)
#t)
#f)))))
(tests:update-central-meta-info run-id test-id new-cpu-load new-disk-free (calc-minutes) #f #f)
(if kill-job?
(begin
(mutex-lock! m)
;; NOTE: The pid can change as different steps are run. Do we need handshaking between this
;; section and the runit section? Or add a loop that tries three times with a 1/4 second
;; between tries?
(let* ((pid1 (launch:einf-pid exit-info)) ;; (vector-ref exit-info 0))
(pid2 (rmt:test-get-top-process-pid run-id test-id))
(pids (delete-duplicates (filter number? (list pid1 pid2)))))
(if (not (null? pids))
(begin
(for-each
(lambda (pid)
(handle-exceptions
exn
(begin
(debug:print-info 0 #f "Unable to kill process with pid " pid ", possibly already killed.")
(debug:print 0 *default-log-port* " message: " ((condition-property-accessor 'exn 'message) exn)))
(debug:print 0 *default-log-port* "WARNING: Request received to kill job " pid) ;; " (attempt # " kill-tries ")")
(debug:print-info 0 #f "Signal mask=" (signal-mask))
;; (if (process:alive? pid)
;; (begin
(map (lambda (pid-num)
(process-signal pid-num signal/term))
(process:get-sub-pids pid))
(thread-sleep! 5)
;; (if (process:process-alive? pid)
(map (lambda (pid-num)
(handle-exceptions
exn
#f
(process-signal pid-num signal/kill)))
(process:get-sub-pids pid))))
;; (debug:print-info 0 #f "not killing process " pid " as it is not alive"))))
pids)
(tests:test-set-status! run-id test-id "KILLED" "KILLED" (args:get-arg "-m") #f))
(begin
(debug:print 0 *default-log-port* "ERROR: Nothing to kill, pid1=" pid1 ", pid2=" pid2)
(tests:test-set-status! run-id test-id "KILLED" "FAILED TO KILL" (args:get-arg "-m") #f)
)))
(mutex-unlock! m)
|
|
|
|
|
|
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
|
df
#f))))
(set! kill-job? (or (test-get-kill-request run-id test-id) ;; run-id test-name itemdat))
(and runtlim (let* ((run-seconds (- (current-seconds) start-seconds))
(time-exceeded (> run-seconds runtlim)))
(if time-exceeded
(begin
(debug:print-info 0 *default-log-port* "KILLING TEST DUE TO TIME LIMIT EXCEEDED! Runtime=" run-seconds " seconds, limit=" runtlim)
#t)
#f)))))
(tests:update-central-meta-info run-id test-id new-cpu-load new-disk-free (calc-minutes) #f #f)
(if kill-job?
(begin
(mutex-lock! m)
;; NOTE: The pid can change as different steps are run. Do we need handshaking between this
;; section and the runit section? Or add a loop that tries three times with a 1/4 second
;; between tries?
(let* ((pid1 (launch:einf-pid exit-info)) ;; (vector-ref exit-info 0))
(pid2 (rmt:test-get-top-process-pid run-id test-id))
(pids (delete-duplicates (filter number? (list pid1 pid2)))))
(if (not (null? pids))
(begin
(for-each
(lambda (pid)
(handle-exceptions
exn
(begin
(debug:print-info 0 *default-log-port* "Unable to kill process with pid " pid ", possibly already killed.")
(debug:print 0 *default-log-port* " message: " ((condition-property-accessor 'exn 'message) exn)))
(debug:print 0 *default-log-port* "WARNING: Request received to kill job " pid) ;; " (attempt # " kill-tries ")")
(debug:print-info 0 *default-log-port* "Signal mask=" (signal-mask))
;; (if (process:alive? pid)
;; (begin
(map (lambda (pid-num)
(process-signal pid-num signal/term))
(process:get-sub-pids pid))
(thread-sleep! 5)
;; (if (process:process-alive? pid)
(map (lambda (pid-num)
(handle-exceptions
exn
#f
(process-signal pid-num signal/kill)))
(process:get-sub-pids pid))))
;; (debug:print-info 0 *default-log-port* "not killing process " pid " as it is not alive"))))
pids)
(tests:test-set-status! run-id test-id "KILLED" "KILLED" (args:get-arg "-m") #f))
(begin
(debug:print 0 *default-log-port* "ERROR: Nothing to kill, pid1=" pid1 ", pid2=" pid2)
(tests:test-set-status! run-id test-id "KILLED" "FAILED TO KILL" (args:get-arg "-m") #f)
)))
(mutex-unlock! m)
|
| ︙ | | | ︙ | |
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
|
(launch:monitor-job run-id test-id item-path fullrunscript ezsteps test-name tconfigreg exit-info m work-area runtlim misc-flags)))
(th1 (make-thread monitorjob "monitor job"))
(th2 (make-thread runit "run job")))
(set! job-thread th2)
(thread-start! th1)
(thread-start! th2)
(thread-join! th2)
(debug:print-info 0 #f "Megatest exectute of test " test-name ", item path " item-path " complete. Notifying the db ...")
(hash-table-set! misc-flags 'keep-going #f)
(thread-join! th1)
(thread-sleep! 1) ;; givbe thread th1 a chance to be done TODO: Verify this is needed. At 0.1 I was getting fail to stop, increased to total of 1.1 sec.
(mutex-lock! m)
(let* ((item-path (item-list->path itemdat))
;; only state and status needed - use lazy routine
(testinfo (rmt:get-testinfo-state-status run-id test-id)))
|
|
|
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
|
(launch:monitor-job run-id test-id item-path fullrunscript ezsteps test-name tconfigreg exit-info m work-area runtlim misc-flags)))
(th1 (make-thread monitorjob "monitor job"))
(th2 (make-thread runit "run job")))
(set! job-thread th2)
(thread-start! th1)
(thread-start! th2)
(thread-join! th2)
(debug:print-info 0 *default-log-port* "Megatest exectute of test " test-name ", item path " item-path " complete. Notifying the db ...")
(hash-table-set! misc-flags 'keep-going #f)
(thread-join! th1)
(thread-sleep! 1) ;; givbe thread th1 a chance to be done TODO: Verify this is needed. At 0.1 I was getting fail to stop, increased to total of 1.1 sec.
(mutex-lock! m)
(let* ((item-path (item-list->path itemdat))
;; only state and status needed - use lazy routine
(testinfo (rmt:get-testinfo-state-status run-id test-id)))
|
| ︙ | | | ︙ | |
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
|
;; if the current status is AUTO the defer to the calculated value but qualify (i.e. make this AUTO-WARN)
(if (equal? (db:test-get-status testinfo) "AUTO") "AUTO-WARN" "WARN"))
((eq? (launch:einf-rollup-status exit-info) 3) "CHECK")
((eq? (launch:einf-rollup-status exit-info) 4) "WAIVED")
((eq? (launch:einf-rollup-status exit-info) 5) "ABORT")
((eq? (launch:einf-rollup-status exit-info) 6) "SKIP")
(else "FAIL")))) ;; (db:test-get-status testinfo)))
(debug:print-info 1 #f "Test exited in state=" (db:test-get-state testinfo) ", setting state/status based on exit code of " (launch:einf-exit-status exit-info) " and rollup-status of " (launch:einf-rollup-status exit-info))
(tests:test-set-status! run-id
test-id
new-state
new-status
(args:get-arg "-m") #f)
;; need to update the top test record if PASS or FAIL and this is a subtest
;; NO NEED TO CALL roll-up-pass-fail-counts HERE, THIS IS DONE IN roll-up-pass-fail-counts called by tests:test-set-status!
|
|
|
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
|
;; if the current status is AUTO the defer to the calculated value but qualify (i.e. make this AUTO-WARN)
(if (equal? (db:test-get-status testinfo) "AUTO") "AUTO-WARN" "WARN"))
((eq? (launch:einf-rollup-status exit-info) 3) "CHECK")
((eq? (launch:einf-rollup-status exit-info) 4) "WAIVED")
((eq? (launch:einf-rollup-status exit-info) 5) "ABORT")
((eq? (launch:einf-rollup-status exit-info) 6) "SKIP")
(else "FAIL")))) ;; (db:test-get-status testinfo)))
(debug:print-info 1 *default-log-port* "Test exited in state=" (db:test-get-state testinfo) ", setting state/status based on exit code of " (launch:einf-exit-status exit-info) " and rollup-status of " (launch:einf-rollup-status exit-info))
(tests:test-set-status! run-id
test-id
new-state
new-status
(args:get-arg "-m") #f)
;; need to update the top test record if PASS or FAIL and this is a subtest
;; NO NEED TO CALL roll-up-pass-fail-counts HERE, THIS IS DONE IN roll-up-pass-fail-counts called by tests:test-set-status!
|
| ︙ | | | ︙ | |
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
|
(args:get-arg ":runname")
(getenv "MT_RUNNAME")))
(fulldir (conc linktree "/"
target "/"
runname)))
(if (and linktree (file-exists? linktree)) ;; can't proceed without linktree
(begin
(debug:print-info 0 #f "Have -run with target=" target ", runname=" runname ", fulldir=" fulldir ", testpatt=" (or (args:get-arg "-testpatt") "%"))
(if (not (file-exists? fulldir))
(create-directory fulldir #t)) ;; need to protect with exception handler
(if (and target
runname
(file-exists? fulldir))
(let ((tmpfile (conc fulldir "/.megatest.cfg." (current-seconds)))
(targfile (conc fulldir "/.megatest.cfg-" megatest-version "-" megatest-fossil-hash))
(rconfig (conc fulldir "/.runconfig." megatest-version "-" megatest-fossil-hash)))
(if (file-exists? rconfig) ;; only cache megatest.config AFTER runconfigs has been cached
(begin
(debug:print-info 0 #f "Caching megatest.config in " tmpfile)
(configf:write-alist *configdat* tmpfile)
(system (conc "ln -sf " tmpfile " " targfile))))
)))
(debug:print-info 1 #f "No linktree yet, no caching configs.")))))
;; gather available information, if legit read configs in this order:
;;
;; if have cache;
;; read it a return it
;; else
|
|
|
|
|
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
|
(args:get-arg ":runname")
(getenv "MT_RUNNAME")))
(fulldir (conc linktree "/"
target "/"
runname)))
(if (and linktree (file-exists? linktree)) ;; can't proceed without linktree
(begin
(debug:print-info 0 *default-log-port* "Have -run with target=" target ", runname=" runname ", fulldir=" fulldir ", testpatt=" (or (args:get-arg "-testpatt") "%"))
(if (not (file-exists? fulldir))
(create-directory fulldir #t)) ;; need to protect with exception handler
(if (and target
runname
(file-exists? fulldir))
(let ((tmpfile (conc fulldir "/.megatest.cfg." (current-seconds)))
(targfile (conc fulldir "/.megatest.cfg-" megatest-version "-" megatest-fossil-hash))
(rconfig (conc fulldir "/.runconfig." megatest-version "-" megatest-fossil-hash)))
(if (file-exists? rconfig) ;; only cache megatest.config AFTER runconfigs has been cached
(begin
(debug:print-info 0 *default-log-port* "Caching megatest.config in " tmpfile)
(configf:write-alist *configdat* tmpfile)
(system (conc "ln -sf " tmpfile " " targfile))))
)))
(debug:print-info 1 *default-log-port* "No linktree yet, no caching configs.")))))
;; gather available information, if legit read configs in this order:
;;
;; if have cache;
;; read it a return it
;; else
|
| ︙ | | | ︙ | |
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
|
;; Now create the link from the test path to the link tree, however
;; if the test is iterated it is necessary to create the parent path
;; to the iteration. use pathname-directory to trim the path by one
;; level
(if (not not-iterated) ;; i.e. iterated
(let ((iterated-parent (pathname-directory (conc lnkpath "/" item-path))))
(debug:print-info 2 #f "Creating iterated parent " iterated-parent)
(handle-exceptions
exn
(begin
(debug:print 0 *default-log-port* "ERROR: Failed to create directory " iterated-parent ((condition-property-accessor 'exn 'message) exn) ", exiting")
(exit 1))
(create-directory iterated-parent #t))))
|
|
|
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
|
;; Now create the link from the test path to the link tree, however
;; if the test is iterated it is necessary to create the parent path
;; to the iteration. use pathname-directory to trim the path by one
;; level
(if (not not-iterated) ;; i.e. iterated
(let ((iterated-parent (pathname-directory (conc lnkpath "/" item-path))))
(debug:print-info 2 *default-log-port* "Creating iterated parent " iterated-parent)
(handle-exceptions
exn
(begin
(debug:print 0 *default-log-port* "ERROR: Failed to create directory " iterated-parent ((condition-property-accessor 'exn 'message) exn) ", exiting")
(exit 1))
(create-directory iterated-parent #t))))
|
| ︙ | | | ︙ | |
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
|
(common:nice-path lnkpath)
lnkpath)
testname "")
;; (rmt:general-call 'test-set-rundir run-id lnkpath testname "") ;; toptest-path)
(if (or (not curr-test-path)
(not (directory-exists? toptest-path)))
(begin
(debug:print-info 2 #f "Creating " toptest-path " and link " lnkpath)
(handle-exceptions
exn
#f ;; don't care to catch and deal with errors here for now.
(create-directory toptest-path #t))
(hash-table-set! *toptest-paths* testname toptest-path)))))
;; The toptest path has been created, the link to the test in the linktree has
|
|
|
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
|
(common:nice-path lnkpath)
lnkpath)
testname "")
;; (rmt:general-call 'test-set-rundir run-id lnkpath testname "") ;; toptest-path)
(if (or (not curr-test-path)
(not (directory-exists? toptest-path)))
(begin
(debug:print-info 2 *default-log-port* "Creating " toptest-path " and link " lnkpath)
(handle-exceptions
exn
#f ;; don't care to catch and deal with errors here for now.
(create-directory toptest-path #t))
(hash-table-set! *toptest-paths* testname toptest-path)))))
;; The toptest path has been created, the link to the test in the linktree has
|
| ︙ | | | ︙ | |
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
|
(if (not remote-megatest)(set! remote-megatest local-megatest)) ;; "megatest"))
(set! mt-bindir-path (pathname-directory remote-megatest))
(if launcher (set! launcher (string-split launcher)))
;; set up the run work area for this test
(if (and (args:get-arg "-preclean") ;; user has requested to preclean for this run
(not (member (db:test-get-rundir testinfo)(list "n/a" "/tmp/badname")))) ;; n/a is a placeholder and thus not a read dir
(begin
(debug:print-info 0 #f "attempting to preclean directory " (db:test-get-rundir testinfo) " for test " test-name "/" item-path)
(runs:remove-test-directory testinfo 'remove-data-only))) ;; remove data only, do not perturb the record
;; prevent overlapping actions - set to LAUNCHED as early as possible
;;
(tests:test-set-status! run-id test-id "LAUNCHED" "n/a" #f #f) ;; (if launch-results launch-results "FAILED"))
(rmt:roll-up-pass-fail-counts run-id test-name item-path #f "LAUNCHED")
(set! diskpath (get-best-disk *configdat* tconfig))
(if diskpath
(let ((dat (create-work-area run-id run-info keyvals test-id test-path diskpath test-name itemdat)))
(set! work-area (car dat))
(set! toptest-work-area (cadr dat))
(debug:print-info 2 #f "Using work area " work-area))
(begin
(set! work-area (conc test-path "/tmp_run"))
(create-directory work-area #t)
(debug:print 0 *default-log-port* "WARNING: No disk work area specified - running in the test directory under tmp_run")))
(set! cmdparms (base64:base64-encode
(z3:encode-buffer
(with-output-to-string
|
|
|
|
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
|
(if (not remote-megatest)(set! remote-megatest local-megatest)) ;; "megatest"))
(set! mt-bindir-path (pathname-directory remote-megatest))
(if launcher (set! launcher (string-split launcher)))
;; set up the run work area for this test
(if (and (args:get-arg "-preclean") ;; user has requested to preclean for this run
(not (member (db:test-get-rundir testinfo)(list "n/a" "/tmp/badname")))) ;; n/a is a placeholder and thus not a read dir
(begin
(debug:print-info 0 *default-log-port* "attempting to preclean directory " (db:test-get-rundir testinfo) " for test " test-name "/" item-path)
(runs:remove-test-directory testinfo 'remove-data-only))) ;; remove data only, do not perturb the record
;; prevent overlapping actions - set to LAUNCHED as early as possible
;;
(tests:test-set-status! run-id test-id "LAUNCHED" "n/a" #f #f) ;; (if launch-results launch-results "FAILED"))
(rmt:roll-up-pass-fail-counts run-id test-name item-path #f "LAUNCHED")
(set! diskpath (get-best-disk *configdat* tconfig))
(if diskpath
(let ((dat (create-work-area run-id run-info keyvals test-id test-path diskpath test-name itemdat)))
(set! work-area (car dat))
(set! toptest-work-area (cadr dat))
(debug:print-info 2 *default-log-port* "Using work area " work-area))
(begin
(set! work-area (conc test-path "/tmp_run"))
(create-directory work-area #t)
(debug:print 0 *default-log-port* "WARNING: No disk work area specified - running in the test directory under tmp_run")))
(set! cmdparms (base64:base64-encode
(z3:encode-buffer
(with-output-to-string
|
| ︙ | | | ︙ | |