I have an issue that from time to time one of the EC2 instances within my cluster have its ECS-agent disconnected. This silently removes the EC2 instance from the cluster (i.e. not eligible to run any services anymore) and silently drains my cluster from serving servers. I have my cluster backed with an autoscaling group, spawning servers to keep up the healthy amount. But the ECS-agent'disconnected servers are not marked as unhealthy, so the AS-group thinks everything is alright.
I have the feeling there must be something (easy) to mitigate this, or I'm having a big issue with choosing ECS and using it in production.
We had this issue for a long time. With each new AWS ECS-optimized AMI it got better, but as of 3 months ago it still happened from time to time. As mcheshier mentioned make sure to always use the latest AMI or at least the latest aws ecs agent
The only way we were able to resolve it was through:
Timed autoscale rotations
We would try to prevent it by scaling up and down at random times
Good cloudwatch alerts
We happened to have our application set up as a bunch of microservices that were all queue (SQS) based. We could scale up and down based on queues. We had decent monitoring set up that let us approximate rates of queues across number of ECS containers. When we detected that the rate was off we would rotate that whole ECS instance. Ie. Say our cluster deployed 4 running containers of worker-1. We approximate that each worker does 1000 messages per 5 minutes. If our queue rate was 3000 per 5 minutes and we had 4 workers, then 1 was not working as expected. We had some scripts set up in lambda to find the faulty one and terminate the entire instance that ran that container.
I hope this helps, I realize it's specific to our in-house application, but the advice I can give you and anyone else is to take the initiative and put as many metrics out there as you can. This will let you do some neat analytics and look for kinks in the system, this being one of them.
Related
I'm using AWS cloudwatch for scaling my application. I created launch configuration, autoscaling group, upscaling and downscaling alarm and policies. The problem is it is taking 5 mins to launch an instance from an AMI. Is there a way to reduce the start-up time from 5 to 2-3 mins?
No, it isn't possible to speed up the provisioning of a new EC2 instance by an AutoScaling scale up action.
I think it's important to appreciate all that EC2 is doing in those 5 minutes. It's building a new virtual machine, installing an image of an operating system on it, hooking it up to a network and bringing it into service. That's pretty impressive for 5 minutes work if you ask me.
If you're needing to scale up that quickly, then quite frankly you're doing it wrong. Even with autoscaling, you should always be a little over provisioned for your expected load. If you start getting close to that over scaled limit, then it's time to autoscale up. Don't provision exactly what you need, it won't work well.
The start up time depends on a few things:
The availability of resources for your instance type within the
availability zone.
The size of the AMI. In the case of a custom AMI image, it may need to be copied to the correct internal storage for the VM.
Initialization procedures. For windows, some images with user-data scripts can require a reboot to join the domain.
This is an old question, and as I have seen. Start times have improved for EC2 over the past years. Some providers like Google Cloud can provide servers in under a minute. So if your workload is that demanding, you may research the available providers and their operational differences.
We run a server architecture where we have an X number of base servers which are always on. Our servers process jobs sent to them and the vast majority of our job requests come in during the workday. To facilitate this particular spike in volume, we use EC2 auto-scaling.
I prefer to launch servers through auto-scaling with as much of a configured AMI as possible as opposed to launching from a base AMI and installing packages through long Chef or Puppet scripts.
In our current build process, we implement changes to our code base late at night when only our base servers are needed and no servers are launched through auto-scaling. But every once in a while, we'll have a critical bug fix that needs to be implemented immediately during the day.
We have a rather large EBS hard drive associated with our servers (app. 400 GB) and AMI creation of a base server with the latest changes usually takes upwards of one hour. This isn't a problem for late night deployments when no additional servers need to be launched, but causes us to lose valuable time during the day because it prevents us from launching additional servers when the latest AMI isn't ready.
Is there anything out there which can speed up the AMI creation process here? I've heard of Netflix's Aminator and Boxfuse, but are there any other alternatives? Also, how do these services stack up against one another?
I have a 23 node cluster running CoreOS Stable 681.2.0 on AWS across 4 availability zones. All nodes are running etcd2 and flannel. Of the 23 nodes, 8 are dedicated etcd2 nodes, the rest are specifically designated as etcd2 proxies.
Scheduled to the cluster are 3 nginx plus containers, a private Docker registry, SkyDNS, and 4 of our application containers. The application containers register themselves with with etcd2 and the nginx containers pick up any changes, render the necessary files, and finally reload.
This all works perfectly, until a singe etcd2 node is unavailable for any reason.
If the cluster of voting etcd2 members loses connectivity to a even a single other voting etcd2 member, all of the services scheduled to the fleet become unstable. Scheduled services begin stopping and starting without my intervention.
As a test, I began stopping the EC2 instances which host voting etcd2 nodes until quorum was lost. After the first etcd2 node was stopped, the above symptoms began. After a second node, services became unstable, with no observable change. Then, after the third was stopped quorum was lost and all units were unscheduled. I then started all three etcd2 nodes again and within 60 seconds the cluster had returned to a stable state.
Subsequent tests yield identical results.
Am I hitting a known bug in etcd2, fleet or CoreOS?
Is there a setting I can modify to keep units scheduled onto a node even if etcd is unavailable for any reason?
I've experienced the same thing. In my case, when I ran 1 specific unit it caused everything to blow up. Scheduled and perfectly fine running units were suddenly lost without any notice, even machines dropping out of the cluster.
I'm still not sure what the exact problem was, but I think it might have had something to do with etcd vs etcd2. I had a dependency of etcd.service in the unit file, which (I think, not sure) caused CoreOS to try and start etcd.service, while etcd2.service was already running. This might have caused the conflict in my case, and messed up the etcd registry of units and machines.
Something similar might be happening to you, so I suggest you check each host whether you're running etcd or etcd2 and check your unit files to see which one they depend on.
I'm writing the backend for a mobile app that does some cpu intensive work. We anticipate the app will not have heavy usage most of the time, but will have occasional spikes of high demand. I was thinking what we should do is reserve a couple of 24/7 servers to handle the steady-state of low demand traffic and then add and remove EC2 instances as needed to handle the spikes. The mobile app will first hit a simple load balancing server that does a simple round-robin user distribution among all the available processing servers. The load balancer will handle bringing new EC2 instances up and turning them back off as needed.
Some questions:
I've never written something like this before, does this sound like a good strategy?
What's the best way to handle bringing new EC2 instances up and back down? I was thinking I could just create X instances ahead of time, set them up as needed (install software, etc), and then stop each instance. The load balancer will then start and stop the instances as needed (eg through boto). I think this should be a lot faster and easier than trying to create new instances and install everything through a script or something. Good idea?
One thing I'm concerned about here is the cost of turning EC2 instances off and back on again. I looked at the AWS Usage Report and had difficulty interpreting it. I could see starting a stopped instance being a potentially costly operation. But it seems like since I'm just starting a stopped instance rather than provisioning a new one from scratch it shouldn't be too bad. Does that sound right?
This is a very reasonable strategy. I used it successfully before.
You may want to look at Elastic Load Balancing (ELB) in combination with Auto Scaling. Conceptually the two should solve this exact problem.
Back when I did this around 2010, ELB had some problems with certain types of HTTP requests that prevented us from using it. I understand those issues are resolved.
Since ELB was not an option, we manually launched instances from EBS snapshots as needed and manually added them to an NGinX load balancer. That certainly could have been automated using the AWS APIs, but our peaks were so predictable (end of month) that we just tasked someone to spin up the new instances and didn't get around to automating the task.
When an instance is stopped, I believe the only cost that you pay is for the EBS storage backing the instance and its data. Unless your instances have a huge amount of data associated, the EBS storage charge should be minimal. Perhaps things have changed since I last used AWS, but I would be surprised if this changed much if at all.
First with regards to costs, whether an instance is started from scratch or from a stopped state has no impact on cost. You are billed for the amount of compute units you use over time, period.
Second, what you are looking to do is called autoscaling. What you do is setup up a launch config that specifies an AMI you are going to use (along with any user-data configs you are using, the ELB and availiabilty zones you are going to use, min and max number of instances, etc. You set up a scaling group using that launch config. Then you set up scaling policies to determine what scaling actions are going to be attached to the group. You then attach cloud watch alarms to each of those policies to trigger the scaling actions.
You don't have servers in reserve that you attach to the ELB or anything like that. Everything is based on creating a single AMI that is used as the template for the servers you need.
You should read up on autoscaling at the link below:
http://aws.amazon.com/autoscaling/
We're running a lightweight web app on a single EC2 server instance, which is fine for our needs, but we're wondering about monitoring and restarting it if it goes down.
We have a separate non-Amazon server we'd like to use to monitor the EC2 and start a fresh instance if necessary and shut down the old one. All our user data is on Elastic Storage, so we're not too worried about losing anything.
I was wondering if anyone has any experience of using EC2 in this way, and in particular of automating the process of starting the new instance? We have no problem creating something from scratch, but it seems like it should be a solved problem, so I was wondering if anyone has any tips, links, scripts, tutorials, etc to share.
Thanks.
You should have a look at puppet and its support for AWS. I would also look at the RightScale AWS library as well as this post about starting a server with the RightScale scripts. You may also find this article on web serving with EC2 useful. I have done something similar to this but without the external monitoring, the node monitored itself and shut down when it was no longer needed then a new one would start up later when there was more work to do.
Couple of points:
You MUST MUST MUST back up your Amazon EBS volume.
They claim "better" reliability, but not 100%, and it's SEVERAL orders of magnitude off of S3's "12 9's" of durability. S3 durability >> EBS durability. That's a fact. EBS supports a "snapshots" feature which backs up your storage efficiently and incrementally to S3. Also, with EBS snapshots, you only pay for the compressed deltas, which is typically far far less than the allocated volume size. In another life, I've sent lost-volume emails to smaller customers like you who "thought" that EBS was "durable" and trusted it with the only copy of a mission-critical database... it's heartbreaking.
Your Q: automating start-up of a new instance
The design path you mention is relatively untraveled; here's why... Lots of companies run redundant "hot-spare" instances where the second instance is booted and running. This allows rapid failover (seconds) in the event of "failure" (could be hardware or software). The issue with a "cold-spare" is that it's harder to keep the machine up to date and ready to pick up where the old box left off. More important, it's tricky to VALIDATE that the spare is capable of successfully recovering your production service. Hardware is more reliable than untested software systems. TEST TEST TEST. If you haven't tested your fail-over, it doesn't work.
The simple automation of starting a new EBS instance is easy, bordering on trivial. It's just a one-line bash script calling the EC2 command-line tools. What's tricky is everything on top of that. Such a solution pretty much implies a fully 100% automated deployment process. And this is all specific to your application. Can your app pull down all the data it needs to run (maybe it's stored in S3?). Can you kill you instance today and boot a new instance with 0.000 manual setup/install steps?
Or, you may be talking about a scenario I'll call "re-instancing an EBS volume":
EC2 box dies (root volume is EBS)
Force detach EBS volume
Boot new EC2 instance with the EBS volume
... That mostly works. The gotchas:
Doesn't protect against EBS failures, either total volume loss or an availability loss
Recovery time is O(minutes) assuming everything works just right
Your services need to be configured to restart automatically. It does no good to bring the box back if Nginx isn't running.
Your DNS routes or other services or whatever need to be ok with the IP-address changing. This can be worked around with ElasticIP.
How are your host SSH keys handled? Same name, new host key can break SSH-based automation when it gets the strong-warning for host-key-changed.
I don't have proof of this (other than seeing it happen once), but I believe that EC2/EBS _already_does_this_ automatically for boot-from-EBS instances
Again, the hard part here is on your plate. Can you stop your production service today and bring it up RELIABLY on a new instance? If so, the EC2 part of the story is really really easy.
As a side point:
All our user data is on Elastic Storage, so we're not too worried about losing anything.
I'd strongly suggest to regularly snapshot your EBS (Elastic Block Storage) to S3 if you are not doing that already.
You can use an autoscale group with a min/max/desired quantity of 1. Place the instance behind an ELB and have the autoscale group be triggered by the ELB healthy node count. This allows you to have built in monitoring by cloudwatch and the ELB health check. Anytime there is an issue the instance be replaced by the autoscale service.
If you have not checked 'Protect against accidental termination' you might want to do so.
Even if you have disabled 'Detailed Monitoring' for your instance you should still see the 'StatusCheckFailed' metric for your instance over which you can configure an alarm (In the CloudWatch dashboard)
Your application (hosted in a different server) should receive the alarm and start the instance using the AWS API (or CLI)
Since you have protected against accidental termination you would never need to spawn a new instance.