[BUG] After logs are retrieved by LogMonitor.exe it crashes without any error message
Closed this issue ยท 8 comments
Describe the bug
After our application publishes logs into IIS, event log and other log files the LogMonitor.exe crashes without any log message.
Using the same application with the LogMonitor.exe version 2.0.2, we don't get any crashes.
To Reproduce
Steps to reproduce the behavior:
- Create a pod with LogMonitor.exe as its entrypoint
- Use the below configuration file for it
{
"LogConfig": {
"sources": [
{
"eventFormatMultiLine": false,
"channels": [
{
"level": "error",
"name": "system"
},
{
"level": "Error",
"name": "application"
},
{
"level": "Verbose",
"name": "Epicor App Server"
}
],
"type": "EventLog",
"startAtOldestRecord": true
},
{
"filter": "*.log",
"includeFileNames": true,
"directory": "c:\\inetpub\\logs",
"includeSubdirectories": true,
"type": "File"
}
]
}
}
- Write some log messages into different locations, like IIS or Event Viewer Logs
- The Pod crashes with the following error:
Last State: Terminated
Reason: Error
Exit Code: -1073741819
Started: Fri, 06 Sep 2024 15:47:20 -0600
Finished: Fri, 06 Sep 2024 15:50:20 -0600
Expected behavior
The pod should not be terminated.
Screenshots
None
Configuration
-Tool: Log Monitor
-Version: 2.1.0
Additional context
I can provide the docker file separately through an email to you.
I can provide memory dumps of the failure, they were collected using the following procdump command:
procdump -ma -accepteula -n 200 -e -w logmonitor.exe \vmJumpHost01\Temp\
Thank you for bringing this to our attention, @israelvaldez. Could you please send the docker file and memory dumps to this email: WindowsContainerGitHubIssues@service.microsoft.com? The team would appreciate them as they try to repro the issue.
Memory dump and docker file sent to that address.
@israelvaldez can you please confirm the version of Log Monitor you are using? Analyzing the LogMonitor.DMP file you shared shows the version as 2.0.0.0
I copied the exe from the container image and it was 2.1.0.
Hey @israelvaldez could you please confirm if this issue still exists in Log Monitor Version 2.1.1. We reverted the change that was causing the failures!
We're looking into deploying 2.1.1 into a test cluster to review if this problem is resolved in there. We'll report back once we get this deployed and it is running with what we find.
We tried using version 2.1.1 and we didn't get any constant crashes, as we used to do.
Thanks for looking into this!