I'm migrating a Django application to Redhat Openshift Online. The application is subject to spikes in demand, so I want to use the Openshift Autoscaling functionality.
To test this, I use Apache JMeter to put a lot of load on the server, to see whether the new gears launch I expect. But I'm encountering bugs with the server scaling up, like my deployment scripts not working as expected, or migrations not occurring correctly on the database. Is there a more convenient way to test the auto-scaling than sending a bunch of requests at the server until haproxy launches a new gear?
You can scale applications up or down using both the web console and the rhc command line tool. You can read more about how to do it here: https://developers.openshift.com/managing-your-applications/scaling.html#managing-application-scaling
Can you provide more details about what scripts/migrations are not working correctly on the newly created gears? You can also feel free to send questions/issues to https://developers.openshift.com/contact.html
Related
I've made a pretty simple web application which has a REST API backend service written in Python/Django and a FE service written in JS/React. Both parts are containerized and can be launched locally via docker-compose. They are situated in separate github repositories, and every time a new tag is pushed to the corresponding repo, an image is built and pushed to the corresponding ecr repo via github actions. Until that point everything works smoothly, but the problem is that, I don't know how to properly automate the deployment process to the test and production environments. The goal is to have those environments as similar as possible.
My current solution for test env is to simply upload docker-compose file to the ec2 instance via github actions, then manually run the docker-compose command, which pulls images from ecr.
Even though it's a simple solution, it's not very scalable or automated and requires some work to be done in order to update an application. The desired flow is to have a manual GitHub action in each repository, which would deploy either BE or FE to the test or prod environment without any need to ssh into the server and do any other manipulations with docker.
I was looking at ECS, but it seems that it's a solution for bigger apps, which need several or more instances to run. I want my app to be used by many users, but I'm not sure, when it will happen. So maybe i should stick to something less complicated than ECS. Are there any simpler solutions, which i am missing? Like Elastic beanstack or something from any other provider?
I will be happy to receive a feedback on anything I wrote in this post, thanks!
I want to set up a large scale scraping service. At the moment I use the scrapyrt api (https://scrapyrt.readthedocs.io/en/latest/api.html#) but I would like to run my spiders asynchronously so my crawls can scale up. I heard that scrapyd was the tool of choice for that, however, I would still like to be able to use scrapyrt for debugging purpose. Is it possible to have scrapyd and scrapyrt running next to each other in the same Docker container and have them callable with their own API.
Also when deploying scrapyd on ecs, I would like to have my scrapyd project deployed on each ecs instance I am using instead of having to deploy them manually with scrapyd-client everyt time a new instance is created. Is it something that is easily doable with scrapyd.
Cheers
I have an application developed in Meteor Framework.
We are planning to move it to AWS withmulti AZ deployment
need Master Slave configuration for the Mongo DB
My question is how to achieve this, i believe mongo db comes bundled in with the Framework itself,
never worked on it so any help will be appriciated.
Thanks
Welcome to Stack Overflow.
Mongo is bundled into the development environment, but not the server.
It is normal to host the database either on a different server of your own, or using a database service (there are many around, such as compose.io, Mongolab etc) So Mongo can be set up for load balancing and scaling independently of the app itself.
We have our production site in Elastic Beanstalk. SNS notifications is good feature to keep us updated about the environment status whenever it changes. But, we want to watch the production environment logs closely.
Our project is a java webapplication, we want to check the status of the production environment from other beanstalk environments i.e., beta and staging environment which are also in the same region and within the same application.
Our goals are to
use aws sdk or other aws tools to get the production beanstalk tomcat logs and display in our beta site on some page.
Run some tool periodically from the Beta environment on Live environment. Which basically does the testing of the sites, i.e., whether all code level mappings are good, if any exceptions then email them.
if we break down the point2 into further more -
We have quartz scheduler to schedule a job at a particular time. We are planning to add some script which test the complete environment periodically. Are there any Beanstalk built in tools that tests the complete site, accessing all URLS and testing the DB to java serialize object classes mappings (hibernate mappings) etc.,
We do use S3 elastic beanstalk bucket to check tomcat logs, but would like to implement the step1 & step2 if possible.
--
Thanks
For Item #1:
I don't recommend using beta and dev to watch production. Instead, here's what I'd do:
Setup Pingdom on all the three environments, so we could have a close eye on uptime
Review the Logging Code. Do you have a explicitly pattern/idiom for exception handling in place? Are your logging functioning?
Setup Papertrail with Logback. Why? You'll have realtime aggregate log tailing for each and every machine you setup a syslog receiver for. For beanstalk-maven-plugin, we are about to release an archetype (see an example 'blank' project created out of it). Even if you're not using, its worth it so see how to use it.
Setup Log Rollout to S3. As it is, the usage is quite useless. I suggest you work into something to import for analysis (or better yet: Export for usage from Hive - Which is something Papertrail Does)
Define your Health Check Code accordingly. Think about what could go wrong, in terms of dependencies
Look / Set up some CloudWatch metrics. If you application is heavy and you're on a t1.micro, which conditions it would spike? Use that at your advantage
That are just a few ideas.
w/r/t Item #2:
I suggest you rethink your structure. I actually dislike the idea of using crontab in elastic beanstalk servers, since its error prone (leader_only? Managing output?). Instead, I use my newer favourite crontab webapp - Jenkins, and set up an integration testing / smoke testing artifact, with only the relevant bits to remotely test the instance. Selenium might help, but I guess if your Services are critical, you might be more happy relying in rest-assured, for instance.
Hope it helps.
I have never actually worked for a company which is deploying a Django App (with a large user base), and am curious about what is the best way to do this.
Right now I am hosting a Django App on EC2. The code for the app is sitting in my github account. I have nginx serving static content, and behind it a single apache server running django + mod_wsgi.
I am trying to figure out what the best practice is for "continuous deployment". Right now, after I have added additional functionality I do the following on EC2:
1) git reset HEAD --hard
2) git pull
3) restart apache
4) restart nginx
I have custom logic in my settings.py file so that if I am running on EC2, debug gets set to False, and my databases switch from sqlite3 (development) to mysql (production).
This seems to be working for me now, but I am wondering what is wrong with this process and how could I improve it.
Thanks
I've worked with systems that use Fabric to deploy to multiple servers
I'm the former lead developer at The Texas Tribune, which is 100% Django. We deployed to EC2 using RightScale. I didn't personally write the deployment scripts, but it allowed us to get new instances into the rotation very, very quickly and scales on-demand. it's not cheap, but was worth every penny in my opinion.
I'd agree with John and say that Fabric is the tool to do this sort of thing comfortably. You probably don't want to configure git to automatically deploy with a post commit hook, but you might want to configure a fabric command to run your test suite locally, and then push to production if it passes.
Many people run separate dev and production settings files, rather than having custom logic in there to detect if it's in a production environment. You can inherit from a unified file, and then override the bits that are different between dev and production. Then you start the server using the production file, rather than relying on a single unified settings.py.
If you're just using apache to host the application, you might benefit from a lighter weight solution. Using fastcgi with nginx would allow you to do away with the overhead of apache entirely. There's also a wsgi module for nginx, but I don't know if it's production ready at this point.
There is one more good way how to manage this. For ubuntu/debian amis it is good to manager versions and do deployemnts by packeging your application into .deb