def setStageFailed(self, index):
"""given an index, sets stage to failed, adds to processed stages array"""
self.stages[index].setFailed()
logger.info("ERROR in Stage " + str(index) + ": " + str(self.stages[index]))
self.processedStages.append(index)
for i in nx.dfs_successor(self.G, index).keys():
self.processedStages.append(i)
def test_successor(self):
assert_equal(nx.dfs_successor(self.G, source=0), {
0: [1],
1: [2],
2: [4],
3: [],
4: [3]
})
def setStageFailed(self, index):
"""given an index, sets stage to failed, adds to processed stages array"""
self.stages[index].setFailed()
logger.info("ERROR in Stage " + str(index) + ": " +
str(self.stages[index]))
# This is something we should also directly report back to the user:
print("\nERROR in Stage %s: %s" %
(str(index), str(self.stages[index])))
print("Logfile for (potentially) more information:\n%s\n" %
self.stages[index].logFile)
sys.stdout.flush()
self.processedStages.append(index)
self.failed_stages += 1
# remove job from currently running processes
self.currently_running_stage.remove(index)
for i in nx.dfs_successor(self.G, index).keys():
self.processedStages.append(i)
def setStageFailed(self, index, clientURI):
# given an index, sets stage to failed, adds to processed stages array
# But... only if this stage has already been retried twice (<- for now static)
# Once in while retrying a stage makes sense, because of some odd I/O
# read write issue (NFS race condition?). At least that's what I think is
# happening, so trying this to see whether it solves the issue.
num_retries = self.stages[index].getNumberOfRetries()
if num_retries < 2:
# without a sleep statement, the stage will be retried within
# a handful of milliseconds, that won't solve anything...
# this sleep command will block the server for a small amount
# of time, but should happen only sporadically
time.sleep(STAGE_RETRY_INTERVAL)
self.removeFromRunning(index, clientURI, new_status=None)
self.stages[index].incrementNumberOfRetries()
logger.info("RETRYING: ERROR in Stage " + str(index) + ": " +
str(self.stages[index]))
logger.info(
"RETRYING: adding this stage back to the runnable queue.")
logger.info("RETRYING: Logfile for Stage " +
str(self.stages[index].logFile))
self.requeue(index)
else:
self.removeFromRunning(index, clientURI, new_status="failed")
logger.info("ERROR in Stage " + str(index) + ": " +
str(self.stages[index]))
# This is something we should also directly report back to the user:
print("\nERROR in Stage %s: %s" %
(str(index), str(self.stages[index])))
print("Logfile for (potentially) more information:\n%s\n" %
self.stages[index].logFile)
sys.stdout.flush()
self.processedStages.append(index)
self.failed_stages += 1
for i in nx.dfs_successor(self.G, index).keys():
self.processedStages.append(i)
def test_successor(self):
assert_equal(nx.dfs_successor(self.G,source=0),
{0: [1], 1: [2], 2: [4], 3: [], 4: [3]})
def children(node):
'''returns all childrens smaller than node.'''
L = []
for i in nx.dfs_successor(G2, node).values():
L.extend(i)
return L
def children(node):
'''returns all childrens smaller than node.'''
L = []
for i in nx.dfs_successor(G2,node).values():
L.extend(i)
return L
