Load Runner transaction prompt error:
Action. C (15): Error-27796: failed to connect to server "ip": [10061] connection refused
Problem:I have encountered a problem where multiple users who use the HTTP protocol send requests to the server during one-time test. Set the duration. The error is:
27796, failed to connect to server 'hostname'; port_ld ': 'reason'. 10048.
AnalysisThe server Load balancer has good performance, Fast Packet sending, and fast response from the server. As a result, the ports of the Server Load balancer are fully occupied before timeout. The above error will occur when all of them are full. Run the netstat-Na command. You can see that many ports are opened. Therefore, the TCP Time Out is adjusted. That is, when the last port is not used, the port is released.
Official troubleshooting:View the troubleshooting tool as follows:
Message code 27796 failed to connect to server 'hostname'; port_ld ': 'reason '. unable to connect to the specified server and port. troubleshooting o try to address the reason provided for the connection failure. o try to access the application with a browser from the injector Machine and from another machine (such as the recording machine). o check that you accurately specified the correct host name And port. O ping the host/port. o check if the server application you are trying to access is Running. O if you used a hostname, check if it was resolved to the correct Address. o check if the server application is listening to the right Port. |
This is not a solution.
Successful solution:
In the load generator registry HKEY_LOCAL_MACHINE/system/CurrentControlSet/services/TCPIP/parameters, there are two key values: tcptimedwaitdelaymaxuserport1, where the default value of tcptimedwaitdelay is 30 s, so here, reduce the value to 5 S (adjusted as needed ). 2. You can also increase the value of maxuserport (if this value is not the maximum value ).
Note: After the registry is modified, restart the computer.
1. Modify Press Registration
Trying to modify the Registry
HKEY_LOCAL_MACHINE \ SYSTEM \ currentcontRolset \ Services \ Tcpip \ ParamEters has the following two key values:
Tcptimedwaitdelay --- 1
Maxuserport ---- 65534
The port wait time is reduced, and the maximum available port is increased.
2. Check whether the number of connections is full on the server where nginx and tomcat are located. Enter
Netstat-an | grep time_wait-WC found about 19000 connections. It seems that there are indeed many time_wait connections! 1. Check whether/EXT/sysctl. conf has net. ipv4.tcp _ tw_recycle = 1.
Net. ipv4.tcp _ tw_reuse = 1. The two configurations have been configured. 2. Modify the LoadRunner configuration. Modify the request timeout values in Run Time Setting of vugen and control respectively. In preferences, click options. Three parameters can be modified at one time,We recommend that you change http-request connect timeout to 1000http-request receieve timeout. We recommend that you change it to 1000 step download timeout to 10000 HTTP Keep Alive Time out. We recommend that you change it to 800.