I have a Topology with Spout S, and 3 bolts - A, B, C.
Bolt A reads from Spout S. Bolt A then splits the data into Bolts B and C (based on some filter). Bolts B and C have their respective data sinks.
If I use Storm's anchoring and anchor the tuple at Bolt A, and then later on Bolt B ack's successfully, but Bolt C does a fail(). Will replaying by Storm at Spout S cause duplication of events at Bolt B and so into the data sink at B?
If so, what is the way to avoid that while still using storm's reliability feature of anchoring?
Storm's anchoring feature only support at-least-once processing and there is no support to handle duplicates in case of failure. Depending on you application semantics, this might be an issues or not.
For example if you do idempotent operation later on, duplicate are not issues (an example of an idempotent operation is updating a key-value store -- if you to two put operation because of a duplicate, the state of the key-value store still be the same).
If you have non idempotent operations and duplicates are an issue, you could try to take care if this by you own -- but this is pretty difficult to get right.
As an alternative you could use Trident API instead of low-level API which does provide exactly-once guarantees.
Or, as last resort, use a different system, that does provide exactly-once semantics out-of-the box.
Related
I want to know how Flink handles branching out the result of one operator to multiple downstream operators. If I have an operator A that is connected to 3 other operators B, C and D in parallel, does Flink keep only one copy of the result and send it out to all 3 operators? Assuming that's the case, is the result removed once it has been sent to all 3 operators or is there a garbage collection process?
I haven't found anything concrete about this topic in the Flink documentation. Any relevant resource regarding this topic is highly appreciated.
Before sending record A to operators B, C, D, FLINK serialise record A into bytes. Then those bytes will be copied into buffer pools for sub-partitions for different operators, so there will be multiple copies of record A.
Record A will be GC later.
I am a beginner of storm. Storm's creator created a very impressive method to check every Bolts in topology, which is using XOR.
But I start wondering why he just not use a counter. When a Bolts is successfully executed, the counter will minus one. So when the counter equal with 0, means the whole task is completetly.
Thanks
I believe one can reason why counters are not only inefficient but an incorrect acker tracker mechanism in an always running topology.
Storm tuple topology in itself can be a complex DAG. When a bolt receives ack from multiple downstream sources, what is it to do with the counters? Should it increment them, should it always decrement them? In what order?
Storm tuples have random message Ids. Counters will be finite. A topology runs forever emitting billions of tuples. How will you map the 673686557th tuple to a counter id? With XOR, you only have a single state to maintain and broadcast.
XOR operations are hardware instructions that execute extremely efficiently. Counters are longs which require huge amounts of storage. They have overflow problems and defeat the original requirement of a solution with a low space overhead.
I am reading up on Apache Storm to evaluate if it is suited for our real time processing needs.
One thing that I couldn't figure out until now is — Where does it store the tuples during the time when next node is not available for processing it. For e.g. Let's say spout A is producing at the speed of 1000 tuples per second, but the next level of bolts(that process spout A output) can only collectively consume at a rate of 500 tuples per second. What happens to the other tuples ? Does it have a disk-based buffer(or something else) to account for this ?
Storm used internal in-memory message queues. Thus, if a bolt cannot keep up processing, the messages are buffered there.
Before Storm 1.0.0 those queues may grow out-of-bound (ie, you get an out-of-memory exception and your worker dies). To protect from data loss, you need to make sure that the spout can re-read the data (see https://storm.apache.org/releases/1.0.0/Guaranteeing-message-processing.html)
You could use "max.spout.pending" parameter, to limit the tuples in-flight to tackle this problem though.
As of Storm 1.0.0, backpressure is supported (see https://storm.apache.org/2016/04/12/storm100-released.html). This allows bolt to notify its upstream producers to "slow down" if a queues grows too large (and speed up again in a queues get empty). In your spout-bolt-example, the spout would slow down to emit messaged in this case.
Typically, Storm spouts read off of some persistent store and track that completion of tuples to determine when it's safe to remove or ack a message in that store. Vanilla Storm itself does not persist tuples. Tuples are replayed from the source in the event of a failure.
I have to agree with others that you should check out Heron. Stream processing frameworks have advanced significantly since the inception of Storm.
I have already read related materials about storm parallel but still keep something unclear. Suppose we take Tweets processing as an example. Generally what we are doing is retrieving tweets streaming, counting numbers of words of each tweets and write the numbers into a local file.
My question is how to understand the value of the parallelism of spouts as well as bolts. Within the function of builder.setSpout and builder.setBolt we can assign the parallel value. But in the case of word counting of tweets is it correct that only one spout should be set? More than one spouts are regarded as copies of the first same spout by which identical tweets flow into several spouts. If that is the case what is the value of setting more than one spouts?
Another unclear point is how to assign works to bolts? Is the parallel mechanism achieve in the way of Storm will find currently available bolts to process a next emitting spout? I revise the basic tweets counting code so the final counting results will be written into a specific directory, however, all results are actually combined in one file on nimbus. Therefore after processing data on supervisors all results will be sent back to nimbus. If this is true what is the communication mechanism between nimbus and supervisors?
I really want to figure out those problems!!! Do appreciate for the help!!
Setting the parallelism for spouts larger than one, required that the user code does different things for different instances. Otherwise (as you mentioned already), data is just sent through the topology twice. For example, you can have a list of ports you want to listen to (or a list of different Kafka topics). Thus, you need to ensure, that different instanced listen to different ports or topics... This can be achieved in open(...) method by looking at topology metadata like own task ID, and dop. As each instance has a unique ID, you can partition your ports/topics such that each instance picks different ports/topics from the overall list.
About parallelism: this depends on the connection pattern you are using when pluging your topology together. For example, using shuffleGrouping results in a round robin distribution of your emitted tuples to the consuming bolt instances. For this case, Storm does not "look" if any bolt instance is available for processing. Tuples are simply transfered and buffered at the receiver if necessary.
Furthermore, Nimbus and Supervisor only exchange meta data. There is not dataflow (ie, flow of tuples) between them.
In some cases like "Kafka's Consumer Group" you have queue behaviour - which means that if one consumer read from the queue, other consumer will read different message from the queue.
This will distribute read load from the queue across all workers.
In those cases you can have multiple spout reading from the queue
I'm using Storm to parse and save data from Kafka. The data comes in as some identifiers and then a map<string,string> of varying size. After some munging the end goal is Cassandra.
Should I send the data as one block of tuples or split up the map and send each piece separately?
A tuple should represent a "unit of work" for the next bolt in the stream. If you think of your map as a single entity that gets processed as a single, albeit complex, object then the map should be emitted as a single tuple. If you want different bolts independently processing different map attributes, then break the map into subsequently processable subsets of attributes and emit multiple tuples.
It depends on the size of the tuple you want to send.
Every tuple you emit in Storm will be taken as a serialized message to transmit from one executor to another. You should also take the performance of Netty and LMAX into consideration, since they are used in the latest version of Storm for Inter-worker communication and Intra-worker communication. That is, settings like
Config.TOPOLOGY_RECEIVER_BUFFER_SIZE
Config.TOPOLOGY_TRANSFER_BUFFER_SIZE
Config.TOPOLOGY_EXECUTOR_RECEIVE_BUFFER_SIZE
Config.TOPOLOGY_EXECUTOR_SEND_BUFFER_SIZE
should be taken into account. You could take a look at Understanding the Internal Message Buffers of Storm for more details.