Megatest

Diff
Login

Differences From Artifact [9e72541023]:

To Artifact [331811ff8d]:


1379
1380
1381
1382
1383
1384
1385
1386



1387
1388




1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
			   (let* ((item-path     (db:test-get-item-path new-test-dat))
				  (test-name     (db:test-get-testname new-test-dat))
				  (run-dir       (db:test-get-rundir new-test-dat))    ;; run dir is from the link tree
				  (real-dir      (if (file-exists? run-dir)
						     (resolve-pathname run-dir)
						     #f))
				  (test-state    (db:test-get-state new-test-dat))
				  (test-fulln    (db:test-get-fullname new-test-dat)))



			     (case action
			       ((remove-runs)




				(debug:print-info 0 "test: " test-name " itest-state: " test-state)
				(if (member test-state (list "RUNNING" "LAUNCHED" "REMOTEHOSTSTART" "KILLREQ"))
				    (begin
				      (if (not (hash-table-ref/default test-retry-time test-fulln #f))
					  (begin
					    ;; want to set to REMOVING BUT CANNOT do it here?
					    (hash-table-set! test-retry-time test-fulln (current-seconds))))
				      (if (> (- (current-seconds)(hash-table-ref test-retry-time test-fulln)) allow-run-time)
					  ;; This test is not in a correct state for cleaning up. Let's try some graceful shutdown steps first
					  ;; Set the test to "KILLREQ" and wait five seconds then try again. Repeat up to five times then give
					  ;; up and blow it away.
					  (begin
					    (debug:print 0 "WARNING: could not gracefully remove test " test-fulln ", tried to kill it to no avail. Forcing state to FAILEDKILL and continuing")
					    (mt:test-set-state-status-by-id (db:test-get-id test) "FAILEDKILL" "n/a" #f)
					    (thread-sleep! 1))
					  (begin
					    (mt:test-set-state-status-by-id (db:test-get-id test) "KILLREQ" "n/a" #f)
					    (thread-sleep! 1)))
				      ;; NOTE: This is suboptimal as the testdata will be used later and the state/status may have changed ...
				      (if (null? tal)
					  (loop new-test-dat tal)
					  (loop (car tal)(append tal (list new-test-dat)))))
				    (begin
				      (mt:test-set-state-status-by-id (db:test-get-id test) "REMOVING" "LOCKED" #f)
				      (debug:print-info 1 "Attempting to remove " (if real-dir (conc " dir " real-dir " and ") "") " link " run-dir)
				      (if (and real-dir 
					       (> (string-length real-dir) 5)
					       (file-exists? real-dir)) ;; bad heuristic but should prevent /tmp /home etc.
					  (begin ;; let* ((realpath (resolve-pathname run-dir)))
					    (debug:print-info 1 "Recursively removing " real-dir)
					    (if (file-exists? real-dir)
						(runs:safe-delete-test-dir real-dir)
						(debug:print 0 "WARNING: test dir " real-dir " appears to not exist or is not readable")))
					  (if real-dir 
					      (debug:print 0 "WARNING: directory " real-dir " does not exist")
					      (debug:print 0 "WARNING: no real directory corrosponding to link " run-dir ", nothing done")))
				      (if (symbolic-link? run-dir)
					  (begin
					    (debug:print-info 1 "Removing symlink " run-dir)
					    (handle-exceptions
					     exn
					     (debug:print 0 "ERROR:  Failed to remove symlink " run-dir ((condition-property-accessor 'exn 'message) exn) ", attempting to continue")
					     (delete-file run-dir)))
					  (if (directory? run-dir)
					      (if (> (directory-fold (lambda (f x)(+ 1 x)) 0 run-dir) 0)
						  (debug:print 0 "WARNING: refusing to remove " run-dir " as it is not empty")
						  (handle-exceptions
						   exn
						   (debug:print 0 "ERROR:  Failed to remove directory " run-dir ((condition-property-accessor 'exn 'message) exn) ", attempting to continue")
						   (delete-directory run-dir)))
					      (if run-dir
						  (debug:print 0 "WARNING: not removing " run-dir " as it either doesn't exist or is not a symlink")
						  (debug:print 0 "NOTE: the run dir for this test is undefined. Test may have already been deleted."))
					      ))
				      ;; Only delete the records *after* removing the directory. If things fail we have a record 
				      (if (not remove-data-only)
				          (cdb:remote-run db:delete-test-records db #f (db:test-get-id test)))
				      (if (not (null? tal))
					  (loop (car tal)(cdr tal))))))
			       ((set-state-status)
				(debug:print-info 2 "new state " (car state-status) ", new status " (cadr state-status))
				(mt:test-set-state-status-by-id (db:test-get-id test) (car state-status)(cadr state-status) #f)
				(if (not (null? tal))
				    (loop (car tal)(cdr tal))))
			       ((run-wait)
				(debug:print-info 2 "still waiting, " (length tests) " tests still running")







|
>
>
>


>
>
>
>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|







1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
			   (let* ((item-path     (db:test-get-item-path new-test-dat))
				  (test-name     (db:test-get-testname new-test-dat))
				  (run-dir       (db:test-get-rundir new-test-dat))    ;; run dir is from the link tree
				  (real-dir      (if (file-exists? run-dir)
						     (resolve-pathname run-dir)
						     #f))
				  (test-state    (db:test-get-state new-test-dat))
				  (test-fulln    (db:test-get-fullname new-test-dat))
				  (uname         (db:test-get-uname    new-test-dat))
				  (toplevel-with-children (and (db:test-get-is-toplevel test)
							       (> (cdb:remote-run db:test-toplevel-num-items db run-id test-name) 0))))
			     (case action
			       ((remove-runs)
				;; if the test is a toplevel-with-children issue an error and do not remove
				(if toplevel-with-children
				    (debug:print 0 "WARNING: skipping removal of " test-fulln " with run-id " run-id " as it has sub tests")
				    (begin
				      (debug:print-info 0 "test: " test-name " itest-state: " test-state)
				      (if (member test-state (list "RUNNING" "LAUNCHED" "REMOTEHOSTSTART" "KILLREQ"))
					  (begin
					    (if (not (hash-table-ref/default test-retry-time test-fulln #f))
						(begin
						  ;; want to set to REMOVING BUT CANNOT do it here?
						  (hash-table-set! test-retry-time test-fulln (current-seconds))))
					    (if (> (- (current-seconds)(hash-table-ref test-retry-time test-fulln)) allow-run-time)
						;; This test is not in a correct state for cleaning up. Let's try some graceful shutdown steps first
						;; Set the test to "KILLREQ" and wait five seconds then try again. Repeat up to five times then give
						;; up and blow it away.
						(begin
						  (debug:print 0 "WARNING: could not gracefully remove test " test-fulln ", tried to kill it to no avail. Forcing state to FAILEDKILL and continuing")
						  (mt:test-set-state-status-by-id (db:test-get-id test) "FAILEDKILL" "n/a" #f)
						  (thread-sleep! 1))
						(begin
						  (mt:test-set-state-status-by-id (db:test-get-id test) "KILLREQ" "n/a" #f)
						  (thread-sleep! 1)))
					    ;; NOTE: This is suboptimal as the testdata will be used later and the state/status may have changed ...
					    (if (null? tal)
						(loop new-test-dat tal)
						(loop (car tal)(append tal (list new-test-dat)))))
					  (begin
					    (mt:test-set-state-status-by-id (db:test-get-id test) "REMOVING" "LOCKED" #f)
					    (debug:print-info 1 "Attempting to remove " (if real-dir (conc " dir " real-dir " and ") "") " link " run-dir)
					    (if (and real-dir 
						     (> (string-length real-dir) 5)
						     (file-exists? real-dir)) ;; bad heuristic but should prevent /tmp /home etc.
						(begin ;; let* ((realpath (resolve-pathname run-dir)))
						  (debug:print-info 1 "Recursively removing " real-dir)
						  (if (file-exists? real-dir)
						      (runs:safe-delete-test-dir real-dir)
						      (debug:print 0 "WARNING: test dir " real-dir " appears to not exist or is not readable")))
						(if real-dir 
						    (debug:print 0 "WARNING: directory " real-dir " does not exist")
						    (debug:print 0 "WARNING: no real directory corrosponding to link " run-dir ", nothing done")))
					    (if (symbolic-link? run-dir)
						(begin
						  (debug:print-info 1 "Removing symlink " run-dir)
						  (handle-exceptions
						   exn
						   (debug:print 0 "ERROR:  Failed to remove symlink " run-dir ((condition-property-accessor 'exn 'message) exn) ", attempting to continue")
						   (delete-file run-dir)))
						(if (directory? run-dir)
						    (if (> (directory-fold (lambda (f x)(+ 1 x)) 0 run-dir) 0)
							(debug:print 0 "WARNING: refusing to remove " run-dir " as it is not empty")
							(handle-exceptions
							 exn
							 (debug:print 0 "ERROR:  Failed to remove directory " run-dir ((condition-property-accessor 'exn 'message) exn) ", attempting to continue")
							 (delete-directory run-dir)))
						    (if run-dir
							(debug:print 0 "WARNING: not removing " run-dir " as it either doesn't exist or is not a symlink")
							(debug:print 0 "NOTE: the run dir for this test is undefined. Test may have already been deleted."))
						    ))
					    ;; Only delete the records *after* removing the directory. If things fail we have a record 
					    (if (not remove-data-only)
						(cdb:remote-run db:delete-test-records db #f (db:test-get-id test)))
					    (if (not (null? tal))
						(loop (car tal)(cdr tal))))))))
			       ((set-state-status)
				(debug:print-info 2 "new state " (car state-status) ", new status " (cadr state-status))
				(mt:test-set-state-status-by-id (db:test-get-id test) (car state-status)(cadr state-status) #f)
				(if (not (null? tal))
				    (loop (car tal)(cdr tal))))
			       ((run-wait)
				(debug:print-info 2 "still waiting, " (length tests) " tests still running")