Currently I'm investigating a setup backed by api-platform with the following goals:
the PHP backend MUST yield minimal resource payloads, thus I do not want to embed relations at all
the PHP backend SHOULD be able to run in alternative runtimes, e.g. Swoole
the webserver should push related resources via HTTP2 Push leveraging the built in vulcain support of the api-platform distribution
I cannot find that many resources about those setups - at least not in such a form that they answer subsequent questions sufficiently.
My starting setup was simply based on the api-platform distribution 2.6.8
So, until now I've learned the following things:
out of the box, the caddy + http2 push setup works with the PHP container being based on php:8.1-fpm-alpine - while caddy is obviously directly using php_fastcgi
when I was fooling around with the currently available cache-handler I was able to get the http cache working but I was struggling to find any information about cache invalidation works. The api-platform docs mostly focus on varnish; there is also only a VarnishPurger shipped in the api-platform core. Wring a custom one should not be that hard if the caddy cache-handler somehow allows BAN requests or something similar - where to find info about that? I see that the handler is based on Souin - but as unfamiliar as I am I have no clue how (and if) Souin supports cache invalidation after all.
when changing the php container to be (in my current testing scenario) based on Swoole then php_fastcgi cannot be used in caddy - instead, I ended up using reverse_proxy (as described in vulcain docs) which basically works and serves proper http responses but does not push any resources requested with Preload headers (as I said, it worked when the PHP backend was based on PHP-FPM). How can I debug what happens here? Caddy does not yield any info about the push handling - nor does the vulcain caddy module
Long story short(er): to sum up my questions
how can I figure out why caddy + vulcain is not working in a reverse_proxy setup?
is the current state of the caddy cache handler functional / supported by the api-platform distribution
how to implement/support BAN requests (or other fine grained cache invalidation) for caddy cache handler?
Souin supports the invalidation using the PURGE HTTP method. I already wrote a PR to set Souin in the api-platform/core project but they are busy with the v3.0 release. Maybe in a near future they'll review and probably merge it, I dunno. But if you use a decorator on the varnish purger and use the code I wrote in the PR, you'll be able to purge automatically the associated endpoints to the base route.
Related
I'm still new to AWS and just following the documentation and asking questions here when I get stuck. Please excuse me if this question sounds really noobish.
So far, I've deployed the following:
EB to deploy my REST API
RDS to deploy my psql database
Lambda functions to handle things like authentication & sending JWTs, uploading images to S3, etc.
I have got my basic back end (no caching (just started learning about redis), etc. set up yet, just the bare bones so far) deployed.
I'm still developing my front end, and have not even thought about how I will be deploying it yet (probably another deployment on EB, since I am using universal react). I am just developing it locally but using my production env variables now so I am hitting my deployed API, etc.
One of the MAJOR things I have no idea on how to do is detecting incoming requests from client side to get the client's location by IP address. This is so that I can return the INITIAL results in your general location just like yelp, foursquare, etc. do when you go to to their sites.
For now, I am just building a web app on desktop so I just want to worry about getting the IP address to get the general area of the user. My use case is something similar to other sites you might have used which provides an INITIAL result set for things in your area (think foursquare or yelp).
Here are my questions:
What would be a good way to do this? I'm thinking of handling this in my front end react universal deployment since it will be a node server with rendered page caching. Is this a terrible idea? It would work something like
(1) request from client comes in
(2) get IP from request and lookup the IP location using some service (still not sure what I'm going to use, have found a few plus a nodejs library called node-geoip). Preferably, I can get the zip code since I am trying to save having to do so many queries by unique locations in my database, and instead return results in the zip code and the front end will show an initial map with the initial results in that zip code.
(3) return to client the rendered page with those location params if it exists, otherwise create it, send it, and cache it.
Is the above a really dumb idea? Maybe you have already done something like this, and could share your wisdom :)
Is there an AWS service which can already handle something like this for me? Perhaps there's some functionality which can already do this.
Thanks.
AGAIN - I apologize if this is long winded. I don't know anyone in real life who can help me and I feel alone :(. I appreciate the help you guys can provide.
There are two parts to this:
Getting the user's IP address. You mentioned you're using 'EB' - I presume you mean AWS ELB (Elastic Load Balancer)? If so, then you need to read the X-Forwarded-For HTTP header in your app code, since otherwise what you'll really detect is the ELB's IP address. X-Forwarded-For contains the user's real IP - or rather, the IP of the end-connection being made (there's no telling if this is really a VPN, Proxy or something else-- but it's as far as you can get with an IP.)
Querying an IP DB that can turn the addr into a location object. There are tons of libraries for you. Assuming you're using Node, you can use node-geoip as you mentioned. Or you can just search 'geoip service' on Google and find managed services, like Telize on Mashape. If you don't want to manage the DB lookup yourself or keep the thing up to date, then a managed service would help.
In either case, it's likely that you'll be doing asynchronous look-ups. In that case, you might want to use async/await to get the user's full object before injecting that into your React props and ultimately rendering it as a HTML string that's sent down to the client.
You could also use a library like redial to decorate your components with data requirements, and return a Promise you can await on to know when you're okay to render.
Since you probably want to enable client routing too (i.e. where the user can click on a route in their browser, and the server isn't touched at all), then you will probably need some way to retrieve the IP address/results based on that IP even when the server isn't involved in the initial render.
For that, you could write a REST service that retrieves the results. Or write a GraphQL back-end that gets the data. It doesn't matter how you write it, since the server will have access to the X-Forwarded-For header and can use that to retrieve the results and send back location-aware data.
FYI, I'm writing a React starter kit (called ReactNow) that uses rxjs for handling async streams. It's not ready yet, but it might help you figure out the code layout that would offer a balanced mix between rendering on the server, and writing universal code that requires some heavy lifting from the server.
I've followed the instructions in this post to configure this but redis status is always showing not connected. I tried to turn on diagnostics on redis cache instance but it doesn't even seem like any requests are getting to the service itself. Any ideas?
After some more investigation it looks like there is a thread that covers the majority of the scenarios on the plug in discussions page... I would read through this and ensure you have things in the properly place as well...
https://wordpress.org/support/topic/enabling-with-predis-and-remote-redis
HTTPS is not supported by the predis library, you must use the non HTTPS endpoint which is 6379.
you can use https for redis cache and wordpress. You need to set it to set WP_REDIS_SCHEME to use TLS. See: https://cloud.accigo.se/blog/how-to-set-up-azure-redis-cache-for-wordpress/
Sometimes after editing the wp-config.php file, when you click on diagnostics, it will show you Redis: not found, but you may need to wait a little and try to Enable Redis object and disable Redis Object few time to get the updated information from the wp-config file.
I have put together a blog post which list down detailed steps to configure Redis Cache with WordPress on Azure.
I hope this will help.
Given a list of URLs, how does one implement the following automated task (assuming windows and ubuntu are the available O/Ses)? Are there existing types of tools that can make implementing this easier or do this out of the box?
log in with already-known credentials
for each specified url
request page from server
wait for page to be returned (no specific max time limit)
if request times out, try again (try x times)
if server replies, or x attempts failed, request next url
end for each
// Note: this is intentionally *not* asynchronous to be nice to the web-server.
Background: I'm implementing a worker tool that will request pages from a web server so the data those pages need to crunch through will be cached for later. The worker doesn't care about the resulting pages' contents, although it might care about HTML status codes. I've considered a phantom/casper/node setup, but am not very familiar with this technology and don't want to reinvent the wheel (even though it would be fun).
You can request pages easily with the http module.
Here's an example.
Some people prefer the request module available in npm.
Here's a link to the github page
If you need more than that, you can use phantomjs.
Here's a link to the github page for bridging node and phantom
However, you could also look for simple cli commands for making requests such as wget or curl.
I've been using Yeoman 0.9 up until yesterday when I decided to use the beta 1.0 release on OS X. I use Yeoman to develop an angular app.
With the Yeoman 0.9-->1.0 migration of my app done and working, I'm now keen to extend the dev server fired up by yeoman grunt server to allow me to make cross-domain calls to an API developed by another team, hosted on another server, where they've already allowed cross-domain calls. Up to now we've been using a fake http backend courtesy of angular.js.
--An aside--
If anyone's reading this looking for a quick solution, we got cross-domain calls working by passing the --disable-web-security option to Chrome from the command line using this approach on OS X with Chrome, but it turns off security for all of Chrome (thumbs down) and we can't get Chrome to fire up as a new instance anyway.
--End aside--
Through some digging I found this Stack Overflow post (via the answer to this post) giving me a pretty good idea on what I need to do to get grunt server allowing cross-domain calls.
It essentially involves adding a middleware component to connect to allow alter the headers so that "Access-Control-Allow-Origin" is set to "*" or whatever one likes.
Knowing nothing about Node.js, I can see the change alluded to in the post needs to be made somewhere in the many files created by Yeoman, but where? I've done some raw string searches for "app.configure" and "connect.listen" within the "node_modules" directory set up by Yeoman but I come up with a number of hits, most of which are from examples bundled with the modules and it isn't clear to me which I should edit. In case it helps, here's a snapshot of the directory structure for my angular app:
If anyone can give me some pointers on where these changes may be made I'd really appreciate it!
Solved sans messing around with node using the approach in this github pull request: http://github.com/angular/angular.js/pull/1454
You can use Apache proxy and connect your REST server with gruntjs(angular.js).
Apache would do this:
proxy / -> gruntjs
proxy /service -> REST server
you would use your application hitting Apache and angular.js application would think that is talking with itself so no cross domain problem.
Here is a great tutorial on how to set this up: http://alfrescoblog.com/2014/06/14/angular-js-activiti-webapp-with-activiti-rest/
I am admittedly a complete noob in all things server, Linux, and websockets. I finally managed to set up a VM running Apache, Tomcat, and Railo that I could connect to and serve up CFM pages, all the while learning UNIX command line navigation, server theory, etc, etc...
Here's my problem -- there is only one Railo websocket extension and it is super rinky-dink (I had to modify the CFC just to get the service to start) but I can't get a websocket connection up (I keep getting "unexpected code 200" in Google Chrome). There is minimal documentation, which is not helpful at all.
Basically, I am trying to do some prototyping for a future project that will use websockets. I like Railo for its speed, security, and excellent ability for very database heavy operations. I am interested in Node, but don't know how to get the same security and DB functionality out of Javascript as I can with CFML.
So I have a couple questions: what are my best options for WebSocket servers? Should I be trying to use Apache and/or Tomcat? People keep saying it's totally not worthwhile to have something like Node.js running the websockets portion and something else doing the heavy lifting behind it -- why is this? I'm more than happy writing WS handlers in whatever language if I can just get a nudge in the right direction, some excellent tutorials (I can't seem to find much in this department), or good feedback on how to, from the ground up, set up my Linux box to handle websockets -- and preferably how to handle both websockets and a robust language like Railo.
The Railo extension works fine for me.
What about submitting some test code so that we can debug it? Of course the websockets projects is very young and in full deployment. So feel free to fork and submit patches or suggestions.
You have plenty of options:
Railo Google Group
https://groups.google.com/forum/?fromgroups#!forum/railo
Github Extension Repository
submit a but in the Railo Jira bugtracker
The main problem of node.js is that it's mono-thread : you won't be able to do background tasks using it and local IO will block your server.
A solution I use is Go. It's very fast, has very good concurrency features and has integrated websocket and json libraries (sample : http://gary.beagledreams.com/page/go-websocket-chat.html). An efficient web application server is made in a few dozens lines of Go. You'll find that there is still much less documentation on internet than for java or even node.js through.
There are a few implementations of websockets in java but as I'm in the process of switching everything I had in java to Go I hadn't tested them. I know I use Google gson for the json encoding in java and it's very good.
The "unexpected code 200" is caused by Railo's web socket server sending an outdated response. They changed the web socket spec and Chrome uses the newer spec.
The problem seems to be caused by chrome & co implementing the new spec, "draft-ietf-hybi-thewebsocketprotocol-17". It requires the server to respond with "HTTP/1.1 101 Switching Protocols" rather than 200 OK.
The solution here would be to either update the Railo web socket extension yourself or use some other solution:
Here is a complete demo of a web socket chat server written in PHP.
http://www.flynsarmy.com/2012/02/php-websocket-chat-application-2-0/
I have used this myself to implement a real-time HTML chat served from an Arch Linux machine that I had lying around. Configuration consisted of simply setting up Apache and PHP then changing the IP address in index.html and in server.php to the external ip address of the server machine.
This flynsarmy demo includes a recent version of PHPWebSocket which is an open source web socket server written entirely in PHP and contained in a single file. The demo hooks into three callbacks: connect, message recieved, and disconnect.
The important thing to note, for me, was that the web socket protocol supports text only, not binary so while extending it for my own chat app I had to implement my own commands to help control the server. Commands in my case looked like this:
!kickusers: username, another_username, a_third_username
My server code would check the first character of all messages for a '!' and if present would treat it as a command. Then I slice up the string to get the command "kickusers" and a list of users to kick. Then I call the appropriate kick function and pass it the array of usernames.
Since my scenario was a chat client this meant that the user could literally type this command into chat and the server would accept and respond to it.
The way all this is deployed on my server is like so:
I have Apache serve the index.html page when the user goes to that location on my server in their browser. The only purpose Apache plays here is to give index.html to the client who requested it.
The index.html page contains html to display the chat and javascript to send and recieve chat to/from the server. Basically, index.html is simply a chat client written in HTML and Javascript and it runs in the browser.
I run server.php via ssh on the server to start up the WEB SOCKET server (totally separate from Apache) which just sits there and handles chat stuff like echoing text to the other connected clients etc.
Though the Arch wiki on installing Apache and PHP is specific to Arch in the way that you install the Apache and PHP packages the sections on configuring Apache and PHP apply to all. I'll save you the google query and give you the link here if you like: https://wiki.archlinux.org/index.php/LAMP
As for prototyping, the reason I gave the link to Flynsarmy's chat demo is because his comments are helpful, he wrote a blog about it, and it comes as a very simple yet complete example of how to do something with web sockets in php.