All required changes have been done to respective files like:
stalecheck=true,
keepalive is checked from HTTP request defaults,
retrycount=1,
hc.parameters file changes,
Socket timeout is 240000
Still we see "java.net.SocketException: Connection reset" in response data however I see the valid requests been passed to Server.
The issue wasnt till we reach 3000 users, worked smoothly till 3000 users.
Connection Reset has a lot of meaning, possible reasons are:
One of the server components is not able to handle load so it closes connections on its side
On JMeter side, check that you running in NON GUI mode and that neither JMeter JVM nor injector machine are overloaded which could explain this. See:
https://jmeter.apache.org/usermanual/get-started.html#non_gui
Related
I am running a performance test on a site using Jmeter. Using a load of up to 100 simultaneous users(Threads) the tests pass perfectly, trying to raise this load to 300 users (Threads) I get the following error:
Non HTTP response code: java.net.SocketException / Non HTTP response message: Connection reset
The error occurs in only 0.68% of requests (out of 2412 requests made by 300 users(Threads) only 2 requests generated this error)
I thought it was the maximum number of connections allowed on my server, I went to my application's webconfig and entered the following information: "Min Pool Size = 5; Max Pool Size = 500;". but still not solve the problem.
Does anyone know what I can do to not generate these errors?
Most probably it indicates a problem with your application, try checking:
application logs
application/web server logs and configuration
underlying operating system logs and networking configuration. also pay attention to number of open ports/sockets/handles (can be checked using either built-in OS monitoring tools or JMeter PerfMon Plugin)
If you're absolutely sure that there is nothing wrong with your test script and application and JMeter is configured to behave exactly like a real browser you can follow instructions from JMeterSocketClosed wiki page
More information: The Mysteries of Connection Close
It is a SpringBoot website and deployed in one Linux server. We use Jmeter to do the load test.
We mock 500 users to visit the webiste index page simultaneously. The index page is very simple html, no database connection,so it is a quite short connection.
After about 2 minutes, Jmeter starts to throw timeout exception as bleow
I guess this is because of website reaching its capacity and running out of connection.
I get one quesiton here, why does website reach its capacity 2 minutes later after Jemter starts. If its TCP connection capacity for this website is 1000, I guess it will reach 1000 very soon after the Jmeter starts, not 2 minutes.
Besides, I see many TCP connections are in TIME_WAIT status in Linux server. I guess this may be related with the connection timeout?
Edit: Someone thinks it is running of port. Someone thinks it is running out of connection. And someone thinks it is running out of processing thread(eg. What does this messge java.net.ConnectException/Connection timed out mean in log.jtl file of Jmeter?). I don't know which one is the exact reason...
Most probably this is due to underlying Linux TCP/IP kernel stack configuration, as per Linux TCP/IP tuning for scalability article:
By default, a connection is supposed to stay in the TIME_WAIT state for twice the msl. Its purpose is to make sure any lost packets that arrive after a connection is closed do not confuse the TCP subsystem (the full details of this are beyond the scope of this article, but ask me if you’d like details). The default msl is 60 seconds, which puts the default TIME_WAIT timeout value at 2 minutes. Which means you’ll run out of available ports if you receive more than about 400 requests a second, or if we look back to how nginx does proxies, this actually translates to 200 requests per second. Not good for scaling.
SO double check timeouts along with maximum number of ports/sockets/files on the Linux server - my expectation is that the aforementioned parameters need to be tuned for high loads.
It's also a good practice to have monitoring of baseline OS health metrics in place (CPU, RAM, Network, Disk, swap usage, etc.). You can use i.e. JMeter PerfMon Plugin or JMeter SSHMon Listener for this.
I am relatively new to JMeter however I have been doing Performance testing for almost a decade.
I am working with a proprietary TCP protocol that sends a keep alive periodically - through the existing TCP connection.
I am struggling to understand how I can fork the JMeter 'thread group' to handle a TCP Keep alive received over the same TCP session.
Any ideas?
Thank you brainstrust!
edit: I'm using the TCPsampler and have read the help page. I'll try to provide some more detail shortly about what's happening and how the protocol is written.
edit2: Unfortunately because it's a propriety protocol I cannot reveal the exact nature of the protocol itself but it's largely irrelevant to the problem I'm facing.
Basically, I use the 1st TCP sampler to 'start/authenticate' the session with the server. This is configured the following options:
1. TCPClient classname: LengthPrefixedBinaryTCPClientImpl (my protocol is implemented this standard way)
2. Re-use connection ON.
3. Close connection OFF.
4. Set NoDelay OFF.
5. SO_Linger: nothing
6. Text to send: my hex code for the protocol (this is correct)
I get the response from the first TCP request and then I want to start interacting, however in the session, the server sends a keep alive mid-stream, so occassionally when I send a request, I get an unexpected keep alive response instead (it's an open stream of data).
This is what I would like to solve.
I attempted to use a recursive test fragment, so that on KeepAlive response, it would send the request again however one cannot recurse the test fragments (it throws a Java error on Run attempt).
I hope this gives more context! Thank you for your patience (I'm a newbie SO user!)
Please check the below options if it helps with you sceario:-
If "Re-use connection" is selected, connections are shared between
Samplers in the same thread, provided that the exact same host name
string and port are used. Different hosts/port combinations will use
different connections, as will different threads. If both of "Re-use
connection" and "Close connection" are selected, the socket will be
closed after running the sampler. On the next sampler, another socket
will be created. You may want to close a socket at the end of each
thread loop.
If an error is detected - or "Re-use connection" is not selected - the
socket is closed. Another socket will be reopened on the next sample.
The following properties can be used to control its operation:
tcp.status.prefix text that precedes a status numbertcp.status.suffix
text that follows a status numbertcp.status.properties name of
property file to convert status codes to messagestcp.handler Name of
TCP Handler class (default TCPClientImpl) - only used if not specified
on the GUI
For more details:-https://jmeter.apache.org/usermanual/component_reference.html#TCP_Sampler
I'm using beanstalkd to offload some work to other machines. The setup is a bit unusual, the server is on the internet (public ip) but the consumers are behind adsl lines on some peoples homes. So there is a linux server as client going out through a dynamic ip and connecting to the server to get a job. It's all PHP and I'm using pheanstalk library.
Everything runs smoothly for some time, but then the adsl changes the IP (every 24h hours the provider forces a disconnect-reconnect) the client just hangs, never to go out of "reserve".
I thought that putting a timeout on the reserve would help it, but it didn't. As it seems, the client issues a command and blocks, it never checks the timeout. It just issues a reserve-with-timeout (instead of a simple reserve) and it is the servers responsibility to return a TIME_OUT as the timeout occurs. The problem is, the connection is broken (but the TCP/IP doesn't know about that yet until any of the sides try to talk to the other side) and if the client blocked reading, it will never return.
The library seems to have support for some kind of timeouts locally (for example when trying to connect to server), but it does not seem to contemplate this scenario.
How could I detect the stale connection and force a reconnect? Is there some kind of keepalive on the protocol (and on the pheanstalk itself)?
Thanks!
You could try to close each connection right after the request is answered and reopen a new connection each time.
There is no close() function but you deleting the Pheanstaly Object with unset($pheanstalk) will close it.
This explanation is quite helpful:
Pheanstalk (PHP client for beanstalk) - how do connections work?
I haven't tried it yet, but I came up with the idea of connecting to the beanstalk server through an SSH tunnel. We can enable the ServerAliveCountMax and ServerAliveInterval options on the tunnel, so that a network or server failure will cause the tunnel to close. This should then cause the pheanstalk client to report an error.
I am testing my application using jmeter tool. I am having 2 EC2 m1.small instances behind an ELB (not autoscaling group), 2 caching nodes huge RDS db (Multi-AZ). My Apache (peforked) is configured with default values like 256 maxclients. Both can handle 256-256 request. Now when jmeter throws 500 request I see connection time out error in one of sampler of the jmeter. Can anyone figure out what the problem is?
Thanks in advance.
What does JMeter throw back?
There are couple of things that could happen.
Connection reset: If this error is thrown out from your JMeter then it means that the server has maxed out and cannot support any more concurrent requests. i.e. the 256 threads allotted are all in use serving other requests and this basically means you have hit your limit on the server.
"Address already in use" exception: These kinds of errors are what you must not get when you do load tests. This basically means that there are no available ports on your system make more requests and all ports are busy. This could happen for a variety of reasons but you could try tweaking system settings like ulimit for linux or if your using a windows box you may want to look at the tcpTimedWaitDelay and corresponding settings to see the average turn around time for the ports to be handed over back into the active pool to be reused for the next connection. This condition is called tcp port exhaustion (http://www.outsystems.com/NetworkForums/ViewTopic.aspx?TopicId=6956&Topic=How-to-tune-the-TCP%2FIP-stack-for-high-volume-of-web-requests)
TO get around this you could also try distributed load testing and/or use timers to ensure that you always have the ports to make new connections.