Hi Team, User is facing the below issue.
There have been two runs done with 24 hours - Soak Test. It has been planned to run for 24 hours, but one failed in 4 hours and another in 21 hours,
What we could be inferred is - at 1:07 AM, there is not request sent out from Blazemeter.
there is no issue with our infrastructure servers but not sure why the requests have been stopped earlier and not completed the full test for 24 hours
Can you please help us determine what the root cause of the test stopping is?
The OPL had run out of disk space.
SaaS BLAZEMETER PERFORMANCE TESTING
using the client's on Prem OPL
You need to collect the following files from the OPL:
admin-artifacts.zip and artifacts.zip archives
https://knowledge.broadcom.com/external/article?articleId=209650
In the artifacts.zip file you may see Java Core file and hs_err_pid*.log files
open the hs_err_pid*.log and look for the siginfo: line.
example:
siginfo: si_signo: 7 (SIGBUS), si_code: 2 (BUS_ADRERR), si_addr: 0x00007f1f178f0218
A simple google search of the error should provide direction:
In the above case the 7 (SIGBUS) is an out of disk space error:
Java VM dies with SIGBUS (0x7) when temp directory is full on linux | Confluence | Atlassian Documentation
You can then use the BlazeMeter API call for private locations first to find the ID of the private location and then to find the host information from the private location.
call number one:
GET /private-locations
GET /private-locations/{id}
the second call will return a lot of information.
Search for hostinfo to find the following section:
"installedVersion": "3.5.91",
"hostInfo": {
"diskSpace": {
"/": {
"present": true,
"freeSpace": 932012032,
"freePercent": 0.9
}
},
That is only ~932M