We are using Apache as the web server, and it proxies requests to Jboss (think Tomcat) Java application server using AJP.
We have logging on for Apache and for our web application in Jboss.
We are seeing, not always but sometimes, cases where the processing time for a request in Jboss is less than half a second, but in the Apache log for the same request it is taking over 8 seconds to complete the request.
I can't even think where to start looking and I have not come up with a good Google search to try and work out why Apache is sitting on the request for so long. Any help appreciated.
Disclaimer: Educated guess taken from my experience with running such setups.
Preface
Apache can be configured to allow only a limited number of connections at the same time. In fact this is a prudent way to configure Apache since every connection uses a certain amount of resources and having no upper limit puts you at risk to run into a situation, where your main memory is exhausted and your server becomes unresponsive.
Resource exhaustion
That being said, Apache is usually configured as shown below, your numbers and modules may be different though. The principle still applies.
<IfModule mpm_prefork_module>
StartServers 5
MinSpareServers 5
MaxSpareServers 10
MaxClients 150
MaxRequestsPerChild 0
</IfModule>
This indicates that Apache can process at most 150 concurrent connections.
If a client initiates the 151'th connection, the operating system kernel tries to forward this connection to the Apache process, but it won't answer any more connections. The kernel then enqueues the connection until another connection is closed by Apache.
The time it takes until the kernel can successfully initiate the connection will look to the user as if the request as such takes longer to complete.
The application-server on the other hand, doesn't know about the delay and received the request only after the connection has been initiated. To the application-server therefore everything looks normal.
If you don't have enough resources to increase the concurrent connections in Apache, consider switching to a more resource-efficient web-proxy, like nginx or Varnish.
I don't think apache is acutally slow in your case. I guess you are using keepalived connections between apache and jboss. Under some circumstances, for example the connector is using blocking IO strategy and mean while the number of apache httpd processes are higher than the number of executor threads configurated in jboss connector. It might cause the jboss container thread being blocked after it served a request. You should post your apache and jboss configurations in order to get more specific answers.
Related
I am running tornado behind apache. I have created proxy server.
ProxyRequests On
ProxyPass /chat/ http://localhost:8888/chat/
This code works great and pass all my requests to tornado and returns the response back to client.
Now, I am using tornado for long polling. Some of the requests which finishes in a short interval of time say less than 1 minute this reverse proxy works fine. But certain long polling requests this gives 502 proxy error. The reason for this proxy error is that Apache can hold long polling request for just one minute(by default). It closes the request and hence proxy error is received.
Now, I modified the directive to
ProxyRequests On
ProxyPass /chat/ http://localhost:8888/chat/ timeout=12000
i.e I changed the default timeout to 12000 seconds.
This is currently working fine for me. Bu this is not the best solution to the issue. Ideally long polling requests can exceed any timeout specified. So my questions are
How to make the timeout infinite ? i.e the request is never closed by Apache.
Please also comment: whether the performance of tornado is degraded by going through Apache as proxy server?
I experienced a similar issue with Nginx and solved it the same way as you did. But I changed the timeout to 1 day as it was sufficiently large in my case.
I think you cannot do away with this. The rationale behind this is that Apache (or any proxy server for that matter) has to maintain its performance, which it clearly can't if it has to hold stale or inactive connections. You'd rather let your proxy server proxy more active connections than inactive connections.
Therefore, there is no way to turn off the ProxyTimeout in Apache or even in Nginx (configured using proxy_read_timeout). So if your proxied server is not sending any response within the timeout, then either your application server is taking too long to respond or there is something wrong with your application server or the client is not request for any response. In the first case, you can make safe estimates to set an appropriate timeout. In the second case, you need to fix your application server. And in the third case, you must gracefully handle the situation on the client and reconnect if required.
Coming to your second question, there shouldn't be any difference other than the latency involved between your Apache and your Tornado server. You can very well expose your Tornado server directly to the world but that will come with a few challenges:
1. More ops work - make sure that Tornado process is always up and running.
2. Proxying and load balancing will become more difficult.
3. Worse security as YOU have written that code instead of thousands of expert contributors. So you should not be thinking of running this server as root every. But you can still sort of safely do the same with Apache or Nginx.
Of course the above problems are solvable, but why solve an already solved problem. :)
My web app, an exploded WAR, is hosted by Apache (static content) and Tomcat (dynamic content) via mod_jk. Optionally, there's an ActiveMQ component of this system, but it's currently not being used.
As I understand, each HTTP request will hit Apache. If it's a dynamic content request, Apache will forward the request to Tomcat via mod_jk. To fulfill this request, Tomcat will start a new thread to do the work.
I'm running the app on a 6-core, 12 GB RAM machine.
Besides using the ActiveMQ component, how can I improve my system's performance? Also, please correct me if I'm misstating how Apache and Tomcat communicate.
while (unhappyWithSitePerformance) {
executeLoadTest();
identifyBiggestBottleneck(); // e.g. what breaks first
fixIdentifiedBottleneck();
}
There is no blank silver bullet to provide. You should make sure your load test simulates realistic user behaviour and define the number of (virtual) users you want your server to handle within given answering time. Then tune your server until your goal is met.
Common parameters to look for are
memory consumption
CPU consumption (e.g. certain algorithms)
I/O saturation - e.g. communication to the database, general HTTP traffic saturating the network adapter
Database or backend answering time - e.g. sometimes you'll have to tune the backend, not the webserver itself.
I have apache setup to proxy requests to backend app servers. There are several backend servers and there is a big rewrite_rules file which is processed for each request to properly proxy all the requests to app servers.
One particular app server sometimes takes a long time to respond back, causing the proxy server to become slow if a lot of requests are being proxied to that particular server, as the requests are waiting to get the response back and eventually timeout after the timeout specified for apache requessts.
I would like to put a timeout for this particular rewrite rule only, which is less than the apache timeout value, so that the resources are quickly free and avoid other problems during peak traffic times.
I tried using
or directive to put ProxyTimeout
ProxyTimeout 30
however this is not allowed.
Appreciate any help from experts on this issue.
Thanks,
I just ran into this issue, and solved it by incrementing the overall timeout directive in httpd.conf. So, change:
#
# Timeout: The number of seconds before receives and sends time out.
#
Timeout 60
to a more appropriate value.
It would be great do be able to do this for individual URLs, but it doesn't appear to be possible due to the lack of responses here and the Apache mailing list.
Summary/Quesiton:
I have Apache running with Prefork MPM, running php. I'm trying to use Apache mod_proxy to create a reverse proxy that I can re-route my requests through, so that I can use Apache to do connection pooling. Example impl:
in httpd.conf:
SSLProxyEngine On
ProxyPass /test_proxy/ https://destination.server.com/ min=1 keepalive=On ttl=120
but when I run my test, which is the following command in a loop:
curl -G 'http://localhost:80/test_proxy/testpage'
it doesn't seem to re-use the connections.
After some further reading, it sounds like I'm not getting connection pool functionality because I'm using the Prefork MPM rather than the Worker MPM. So each time I make a request to the proxy, it spins up a new process with its own connection pool (of size one), instead of using the single worker that maintains its own pool. Is that interpretation right?
Background info:
There's an external server that I make requests to, over https, for every page hit on a site that I run.
Negotiating the SSL handshake is getting costly, because I use php and it doesn't seem to support connection pooling - if I get 300 page requests to my site, they have to do 300 SSL handshakes to the external server, because the connections get closed after each script finishes running.
So I'm attempting to use a reverse proxy under Apache to function as a connection pool, to persist the connections across php processes so I don't have to do the SSL handshake as often.
Sources that gave me this idea:
http://httpd.apache.org/docs/current/mod/mod_proxy.html
http://geeksnotes.livejournal.com/21264.html
First of all, your test method cannot demonstrate connection pooling since for every call, a curl client is born and then it dies. Like dead people don't talk a lot, a dead process cannot keep a connection alive.
You have clients that bothers your proxy server.
Client ====== (A) =====> ProxyServer
Let's call this connection A. Your proxy server does nothing, it is just a show off. The handsome and hardworking server is so humble that he hides behind.
Client ====== (A) =====> ProxyServer ====== (B) =====> WebServer
Here, if I am not wrong, the secured connection is A, not B, right?
Repeating my first point, on your test, you are creating a separate client for each request. Every client needs a separate connection. Connection is something that happens between at least two parties. One side leaves and connection is lost.
Okay, let's forget curl now and look together at what we really want to do.
We want to have SSL on A and we want A side of traffic to be as fast as possible. For this aim, we have already separated side B so it will not make A even slower, right?
Connection pooling? There is no such thing as connection pooling at A. Every client comes and goes making a lot of noise. Only thing that can help you to reduce this noise is "Keep-Alive" which means, keeping connection alive from a client for some short period of time so this very same client can ask for other files that will be required by this request. When we are done, we are done.
For connections on B, connections will be pooled; but this will not bring you any performance since on one-server setup you did not have this part of the noise production.
How do we help this system run faster?
If these two servers are on the same machine, we should get rid of the show-off server and continue with our hardworking webserver. It adds a lot of unnecessary work to the system.
If these are separate machines, then you are being nice to web server by taking at least encyrption (for ssl) load from this poor guy. However, you can be even nicer.
If you want to continue on Apache, switch to mpm_worker from mpm_prefork. In case of 300+ concurrent requests, this will work much better. I really have no idea about the capacity of your hardware; but if handling 300 requests is difficult, I believe this little change will help your system a lot.
If you want to have an even more lightweight system, consider nginx as an alternative to Apache. It is very easy to setup to work with PHP and it will have a better performance.
Other than front-end side of things, also consider checking your database server. Connection pooling will make real difference here. Be sure if your PHP installation is configured to reuse connections to database.
In addition, if you are hosting static files on the same system, then move them out either on another web server or do even better by moving static files to a cloud system with CDN like AWS's S3+CloudFront or Rackspace's CloudFiles. Even without CloudFront, S3 will make you happy. Rackspace's solution comes with Akamai!
Taking out static files will make your web server "oh what happened, what is this silence? ohhh heaven!" since you mentioned this is a website and web pages have many static files for each dynamically generated html page most of the time.
I hope you can save the poor guy from the killer work.
Prefork can still pool 1 connection per backend server per process.
Prefork doesn't necessarily create a new process for each frontend request, the server processes are "pooled" themselves and the behavior depends on e.g. MinSpareServers/MaxSpareServers and friends.
To maximise how often a prefork process will have a backend connection for you, avoid very high or low maxspareservers or very high minspareservers as these will result in "fresh" processes acceptin new connections.
You can log %P in your LogFormat directive to help get an idea if how often processes are being reused.
The Problem in my case was, the the connection pooling between reverse proxy and backend server was not taking place because of the Backend Server Apache closing the SSL connection at the end of each HTTPS request.
The backend Apache Server was doing this becuse of the following Directive being present in the httpd.conf:
SetEnvIf User-Agent ".*MSIE.*" nokeepalive ssl-unclean-shutdown
This directive does not make sense when the backend server is connected via a reverse proxy and this can be removed from the backend server config.
I'm writing some code on a mobile device that uses a REST service to retrieve data from a host. That REST services is being proxied by Apache. In test mode I would like to be able to simulate network outages (as if the device has lost it's cell connection) to test the applications handling of intermittent failures. I also need to validate it's behavior with slow network connections.
I'm currently using Traffic Shaper XP to slow the network connection, but now I need something to make the Apache server send connection resets both randomly and on predefined sequences (to setup and repeat specific test scenarios).
I highly recommend https://github.com/Shopify/toxiproxy from Shopify:
Download https://github.com/Shopify/toxiproxy/releases the cli and server
Run the server:
./toxiproxy-server-linux-amd64
On the cli setup proxy to apache on another port e.g. 8080
./toxiproxy-cli create apache -l localhost:8080 -u localhost:80
Make connection slow and unreliable:
./toxiproxy-cli toxic add apache -t latency -a latency=3000
./toxiproxy-cli toxic add apache -t limit_data -a bytes=1000 --tox=0.01
here add 3 second of latency and stop after 1000 bytes for 1% of requests there are other options for bandwidth etc. You can add or remove these during use. Lots of other features and libraries there.
In Apache2 you can make it slow by adjust prefork settings in apache2.conf. The settings below ought to make apache pretty fn slow. They made my local web application take 700% longer to load.
<IfModule mpm_prefork_module>
StartServers 2
MinSpareServers 2
MaxSpareServers 2
MaxClients 4
MaxRequestsPerChild 0
</IfModule>
It looks like DummyNet is the closest thing, but it’s still not quite there. For repeatable testing it would be good to have some control over dropped packets and resets.
Write a little proxy that forwards TCP connections from your app to the apache server and that you can set up in your test to cut the connection after x number of bytes or milliseconds.
On a different (or on the same) computer use the commandline tool ab to get some load on the apache. More informations here.
Is this a Unix or Linux environment? nice it up to give it lower priority then run a high CPU usage task like listening to music, playing a movie, calculating pi, etc. The low priority for Apache should create problems similar to what you're looking for.