I am planning to introduce Hazelcast IQueue as a distributed task queue for the project I am working on.
There are multiple stateless services will consume tasks from such queue, a question come to my mind is ---- load balancing issue.
How can I guarantee the load is relative balanced, while those stateless services consume tasks from the queue?
Queues provide fairness.
Assuming the code is a loop of IQueue.take() and process, then if take() requests come from instance A, instance B, instance A and instance A again, then instance A gets 3 messages and B gets 1.
If this is due to the messages requiring different amounts of time to process, then with sufficient messages this should even out. If instance 1 is on faster hardware than instance 2 then it won't.
If you need the instances to follow a round-robin approach then you'd need to implement this yourself.
You could use something like IAtomicLong to record whose turn is next. All processing stops if the instance whose turn is next goes offline, which might not be what you want. The CP Subsystem provides a fuller featured approach that you can use for scheduling.
Related
I have been researching how to efficiently solve the following use case and I am struggling to find the best solution.
Basically I have a Node.js REST API which handles requests for users from a mobile application. We want some requests to launch background tasks outside of the req/res flow because they are CPU intensive or might just take a while to execute. We are trying to implement or use any existing frameworks which are able to handle different job queues in the following way (or at least compatible with the use case):
Every user has their own set job queues (there are different kind of jobs).
The jobs within one specific queue have to be executed sequentially and only one job at a time but everything else can be executed in parallel (it would be preferable if there are no queues hogging the workers or whatever is actually consuming the tasks so all queues get more or less the same priority).
Some queues might fill up with hundreds of tasks at a given time but most likely they will be empty a lot of the time.
Queues need to be persistent.
We currently have a solution with RabbitMQ with one queue for every kind of task which all the users share. The users dump tasks into the same queues which results in them filling up with tasks from a specific user for a long time and having the rest of users wait for those tasks to be done before their own start to be consumed. We have looked into priority queues but we don't think that's the way to go for our own use case.
The first somewhat logical solution we thought of is to create temporary queues whenever a user needs to run background jobs and have them be deleted when empty. Nevertheless we are not sure if having that many queues is scalable and we are also struggling with dynamically creating RabbitMQ queues, exchanges, etc (we have even read somewhere that it might be an anti-pattern?).
We have been doing some more research and maybe the way to go would be with other stuff such as Kafka or Redis based stuff like BullMQ or similar.
What would you recommend?
If you're on AWS, have you considered SQS? There is no limit on number of standard queues created, and in flight messages can reach up to 120k. This would seem to satisfy your requirements above.
While the mentioned SQS solution did prove to be very scalable our amount of polling we would need to do or use of SNS did not make the solution optimal. On the other hand implementing a self made solution via database polling was too much for our use case and we did not have the time or computational resources to consider a new database in our stack.
Luckily, we ended up finding that the Pro version of BullMQ does have a "Group" functionality which performs a round robin strategy for different tasks within a single queue. This ended up adjusting perfectly to our use case and is what we ended up using.
I have following task to implement using AWS stack:
One job is triggered periodically and put message to queue (SQS). Worker recieves this task and based on it additional tasks need to be created (approximately 1-10 K tasks). And all these tasks are also put to another queue and there are additional workers to process these tasks.
These flow can be described displayed in following way:
Periodic task ->SQS->woker_1(creates more tasks) -> SQS -> workers_2
Based on project conventions and bureaucracy it will take some time to create two separate services for worker_1 that listen to periodic task and creates fine grained tasks and for workers_2 that just process particular tasks, make docker images, CI jobs etc... and get deploy it.
So, here is the tradeof:
1. Spend additional time and create two separate services. On the other hand these services might be really simple. And even there is a doubt to have 2 separate projects.
2. Make this as a one service that put messages to the same queue and also will listen to the messages on the same queue and perorm work for: worker_1 and worker_2.
Any suggestions or thoughts are appreciated!
I don't think there can be a "correct" answer to this, you already have a good list of pros and cons for both options. Some additional things I thought of:
SQS queues don't really allow you to pick out specific types of messages, you pretty much need to read everything first-in-first-out. So if you share queues, you may have less control of prioritizing messages.
For the two services to interact, they need a shared message definition. Sharing the same codebase would make it easier to dev and test the messaging code. Of course, it could also be a shared library.
Deploying both worker types in the same server/application would share resources, which might be more economical at the low end, or it might be confusing at high scale.
It may be possible to develop all the code into the same application, and leave the decision to deployment-time if it is all on the same server and queue or separate servers reading from separate queues. This seems ideal to me.
I'm running a Windows Azure web role which, on most days, receives very low traffic, but there are some (foreseeable) events which can lead to a high amount of background work which has to be done. The background work consists of many database calls (Azure SQL) and HTTP calls to external web services, so it is not really CPU-intensive, but it requires a lot of threads which are waiting for the database or the web service to answer. The background work is triggered by a normal HTTP request to the web role.
I see two options to orchestrate this, and I'm not sure which one is better.
Option 1, Threads: When the request for the background work comes in, the web role starts as many threads as necessary (or queues the individual work items to the thread pool). In this option, I would configure a larger instance during the heavy workload, because these threads could require a lot of memory.
Option 2, Self-Invoking: When the request for the background work comes in, the web role which receives it generates a HTTP request to itself for every item of background work. In this option, I could configure several web role instances, because the load balancer of Windows Azure balances the HTTP requests across the instances.
Option 1 is somewhat more straightforward, but it has the disadvantage that only one instance can process the background work. If I want more than one Azure instance to participate in the background work, I don't see any other option than sending HTTP requests from the role to itself, so that the load balancer can delegate some of the work to the other instances.
Maybe there are other options?
EDIT: Some more thoughts about option 2: When the request for the background work comes in, the instance that receives it would save the work to be done in some kind of queue (either Windows Azure Queues or some SQL table which works as a task queue). Then, it would generate a lot of HTTP requests to itself, so that the load balancer 'activates' all of the role instances. Each instance then dequeues a task from the queue and performs the task, then fetches the next task etc. until all tasks are done. It's like occasionally using the web role as a worker role.
I'm aware this approach has a smelly air (abusing web roles as worker roles, HTTP requests to the same web role), but I don't see the real disadvantages.
EDIT 2: I see that I should have elaborated a little bit more about the exact circumstances of the app:
The app needs to do some small tasks all the time. These tasks usually don't take more than 1-10 seconds, and they don't require a lot of CPU work. On normal days, we have only 50-100 tasks to be done, but on 'special days' (New Year is one of them), they could go into several 10'000 tasks which have to be done inside of a 1-2 hour window. The tasks are done in a web role, and we have a Cron Job which initiates the tasks every minute. So, every minute the web role receives a request to process new tasks, so it checks which tasks have to be processed, adds them to some sort of queue (currently it's an SQL table with an UPDATE with OUTPUT INSERTED, but we intend to switch to Azure Queues sometime). Currently, the same instance processes the tasks immediately after queueing them, but this won't scale, since the serial processing of several 10'000 tasks takes too long. That's the reason why we're looking for a mechanism to broadcast the event "tasks are available" from the initial instance to the others.
Have you considered using Queues for distribution of work? You can put the "tasks" which needs to be processed in queue and then distribute the work to many worker processes.
The problem I see with approach 1 is that I see this as a "Scale Up" pattern and not "Scale Out" pattern. By deploying many small VM instances instead of one large instance will give you more scalability + availability IMHO. Furthermore you mentioned that your jobs are not CPU intensive. If you consider X-Small instance, for the cost of 1 Small instance ($0.12 / hour), you can deploy 6 X-Small instances ($0.02 / hour) and likewise for the cost of 1 Large instance ($0.48) you could deploy 24 X-Small instances.
Furthermore it's easy to scale in case of a "Scale Out" pattern as you just add or remove instances. In case of "Scale Up" (or "Scale Down") pattern since you're changing the VM Size, you would end up redeploying the package.
Sorry, if I went a bit tangential :) Hope this helps.
I agree with Gaurav and others to consider one of the Azure Queue options. This is really a convenient pattern for cleanly separating concerns while also smoothing out the load.
This basic Queue-Centric Workflow (QCW) pattern has the work request placed on a queue in the handling of the Web Role's HTTP request (the mechanism that triggers the work, apparently done via a cron job that invokes wget). Then the IIS web server in the Web Role goes on doing what it does best: handling HTTP requests. It does not require any support from a load balancer.
The Web Role needs to accept requests as fast as they come (then enqueues a message for each), but the dequeue part is a pull so the load can easily be tuned for available capacity (or capacity tuned for the load! this is the cloud!). You can choose to handle these one at a time, two at a time, or N at a time: whatever your testing (sizing exercise) tells you is the right fit for the size VM you deploy.
As you probably also are aware, the RoleEntryPoint::Run method on the Web Role can also be implemented to do work continually. The default implementation on the Web Role essentially just sleeps forever, but you could implement an infinite loop to query the queue to remove work and process it (and don't forget to Sleep whenever no messages are available from the queue! failure to do so will cause a money leak and may get you throttled). As Gaurav mentions, there are some other considerations in robustly implementing this QCW pattern (what happens if my node fails, or if there's a bad ("poison") message, bug in my code, etc.), but your use case does not seem overly concerned with this since the next kick from the cron job apparently would account for any (rare, but possible) failures in the infrastructure and perhaps assumes no fatal bugs (so you can't get stuck with poison messages), etc.
Decoupling placing items on the queue from processing items from the queue is really a logical design point. By this I mean you could change this at any time and move the processing side (the code pulling from the queue) to another application tier (a service tier) rather easily without breaking any part of the essential design. This gives a lot of flexibility. You could even run everything on a single Web Role node (or two if you need the SLA - not sure you do based on some of your comments) most of the time (two-tier), then go three-tier as needed by adding a bunch of processing VMs, such as for the New Year.
The number of processing nodes could also be adjusted dynamically based on signals from the environment - for example, if the queue length is growing or above some threshold, add more processing nodes. This is the cloud and this machinery can be fully automated.
Now getting more speculative since I don't really know much about your app...
By using the Run method mentioned earlier, you might be able to eliminate the cron job as well and do that work in that infinite loop; this depends on complexity of cron scheduling of course. Or you could also possibly even eliminate the entire Web tier (the Web Role) by having your cron job place work request items directly on the queue (perhaps using one of the SDKs). You still need code to process the requests, which could of course still be your Web Role, but at that point could just as easily use a Worker Role.
[Adding as a separate answer to avoid SO telling me to switch to chat mode + bypass comments length limitation] & thinking out loud :)
I see your point. Basically through HTTP request, you're kind of broadcasting the availability of a new task to be processed to other instances.
So if I understand correctly, when an instance receives request for the task to be processed, it pushes that request in some kind of queue (like you mentioned it could either be Windows Azure Queues [personally I would actually prefer that] or SQL Azure database [Not prefer that because you would have to implement your own message locking algorithm]) and then broadcast a message to all instances that some work needs to be done. Remaining instances (or may be the instance which is broadcasting it) can then see if they're free to process that task. One instance depending on its availability can then fetch the task from the queue and start processing that task.
Assuming you used Windows Azure Queues, when an instance fetched the message, it becomes unavailable to other instances immediately for some amount of time (visibility timeout period of Azure queues) thus avoiding duplicate processing of the task. If the task is processed successfully, the instance working on that task can delete the message.
If for some reason, the task is not processed, it will automatically reappear in the queue after visibility timeout period has expired. This however leads to another problem. Since your instances look for tasks based on a trigger (generating HTTP request) rather than polling, how will you ensure that all tasks get done? Assuming you get to process just one task and one task only and it fails since you didn't get a request to process the 2nd task, the 1st task will never get processed again. Obviously it won't happen in practical situation but something you might want to think about.
Does this make sense?
i would definitely go for a scale out solution: less complex, more manageable and better in pricing. Plus you have a lesser risk on downtime in case of deployment failure (of course the mechanism of fault and upgrade domains should cover that, but nevertheless). so for that matter i completely back Gaurav on this one!
Background
I am trying to work out the best structure for an Azure application. Each of my worker roles will spin up multiple long-running jobs. Over time I can transfer jobs from one instance to another by switching them to a readonly mode on the source instance, spinning them up on the target instance, and then spinning the original down on the source instance.
If I have too many jobs then I can tell Azure to spin up extra role instance, and use them for new jobs. Conversely if my load drops (e.g. during the night) then I can consolidate outstanding jobs to a few machines and tell Azure to give me fewer instances.
The trouble is that (as I understand it) Azure provides no mechanism to allow me to decide which instance to stop. Thus I cannot know which servers to consolidate onto, and some of my jobs will die when their instance stops, causing delays for users while I restart those jobs on surviving instances.
Idea 1: I decide which instance to stop, and return from its Run(). I then tell Azure to reduce my instance count by one, and hope it concludes that the broken instance is a good candidate. Has anyone tried anything like this?
Idea 2: I predefine a whole bunch of different worker roles, with identical contents. I can individually stop and start them by switching their instance count from zero to one, and back again. I think this idea would work, but I don't like it because it seems to go against the natural Azure way of doing things, and because it involves me in a lot of extra bookkeeping to manage the extra worker roles.
Idea 3: Live with it.
Any better ideas?
In response to your ideas
Idea 1: I haven't tried doing exactly what you're describing, but in my experience your first instance has a name that ends with _0, the next _1 and I'm sure you can guess the rest. When you decrease the instance count it drops off the instance with the highest number suffix. I would be surprised if it took into account the state of any particular instance.
Idea 2: As I think you hint at, this will create management problems. You can only have 5 different workers per hosted service, so you'll need a service for each group of 5 roles that you want to be able to scale to. Also when you deploy updates you'll have to upload X times more services where X is the maximum number of instances you currently support.
Idea 3: Technically the easiest. Pending some clarification, this is probably what I'd be doing for now. To reduce the downsides of this option it may pay to investigate ways of loading the data faster. There is usually a Goldilocks level (not too much, not too little) of parallelism that helps with this.
You're right - you cannot choose which instance to stop. In general, you'd run the same jobs on each worker role instance, where each instance watches the same queue (or maybe multiple threads or jobs watching multiple queues).
If you really need to run a job on one instance (such as a scheduler), consider using blob leases as the way to constrain this. Create a blob as a mutex. Then, as each instance spins up, the scheduler job attempts to obtain a write lease on that blob. If it succeeds, it runs. If it fails, it simply sleeps (maybe for a minute) and tries again. At some point in the future, as you scale down in instance count, let's say the instance running the scheduler is killed. A minute later (or whatever time span you choose), another instance tries to acquire the lease, succeeds, and now runs the scheduler code.
I have several WorkerRole that only do job for a short time, and it would be a waste of money to put them in a single instance each. We could merge them in a single one, but it'd be a mess and in the far future they are supposed to work independently when the load increases.
Is there a way to create a "multi role" WorkerRole in the same way you can create a "multi site" WebRole?
In negative case, I think I can create a "master worker role", that is able to load the assemblies from a given folder, look for RoleEntryPoint derivated classes with reflection, create instances and invoke the .Run() or .OnStart() method. This "master worker role" will also rethrown unexpected exceptions, and call .OnStop() in all sub RoleEntryPoints when .OnStop() is called in the master one. Would it work? What should I be aware of?
As mentioned by others, this is a very common technique for maximizing utilization of your instances. There may examples and "frameworks" that abstract the worker infrastructure and the actual work you want to be done, including one in this (our) sample: http://msdn.microsoft.com/en-us/library/ff966483.aspx (scroll down to "inside the implementation")
Te most common ways of triggering work are:
Time scheduled workers (like "cron"
jobs)
Message baseds workers (work triggered by the presence of a message).
The code sample mentioned above implements further abstractions for #2 and is easily extensible for #1.
Bear in mind though that all interactions with queues are based on polling. The worker will not wake up with a new message on the queue. You need to actively query the queue for new messages. Querying too often will make Microsoft happy, but probably not you :-). Each query counts as a transaction that is billed (10K of those = $0.01). A good practice is to poll the queue for messages with some kind of delayed back-off. Also, get messages in batches.
Finally, taking this to an extreme, you can also combine web roles and worker roles in a single instance. See here for an example: http://blog.smarx.com/posts/web-page-image-capture-in-windows-azure
Multiple worker roles provide a very clean implementation. However, the cost footprint for idle role instances is going to be much higher than a single worker role.
Role-combining is a common pattern I've seen, working with ISV's on their Windows Azure deployments. You can have a background thread that wakes up every so often and runs a process. Another common implementation technique is to use an Azure Queue to send a message representing a process to execute. You can have multiple queues if you want, or a single command queue. In any case, you would have a queue listener running in a background thread, which would run in each instance. The first one to get the message processes it. You could take it further, and have a timed process pushing those messages onto the queue (maybe every 24 hours, or every hour).
Aside from CPU and memory limits, just remember that a single role can only have a maximum of 5 endpoints (less if you're using Remote Desktop).
EDIT: As of September 2011, role configuration has become much more flexible, now that you have 25 Input endpoints (accessible from the outside world) and 25 Internal endpoints (used for communication between roles) across an entire deployment. The MSDN article is here
I recently blogged about overloading a Web Role, which is somewhat related.
While there's no real issue with the solutions that have been pointed out for finding ways to do multiple worker components within a single Worker Role, I just want you to keep in mind the entire point of having distinct Worker Roles defined in the first place is isolation in the face of faults. If you just shove everything into a single Worker Role instance, just one of those worker components behaving badly has the ability to take down every other worker component in that role. Now all of a sudden you're writing a lot of infrastructure to provide isolation and fault tolerance across components which is pretty much what Azure is there to provide for you.
Again, I'm not saying it's an absolute to strickly do one thing. There's a place where multiple components under a single Worker Role makes sense (especially monaterily). Simply saying that you should keep in mind why it's designed this way in the first place and factor that in appropriately as you plan your architecture.
Why would a 'multi role' be a mess? You could write each worker role implementation as a loosely coupled component and then compose a Worker Role from all appropriate components.
When you later need to separate some of the responsibilities out to a separate worker role, you can compose a new worker role with only this component, while at the same time removing it from the old worker role.
If you wanted to, you could employ late binding so that this could even be done without recompilation, but often I don't think that would be worth the effort.