Core in processor plugin

The following piece of configuration makes Telegraf to core. Latest CentOS 8 and Telegraf. The same configuration works in 1.13.


	unsigned  =  [

telegraf_core.txt (1.9 KB)

Is there any update? I manage to replicate the issue on CentOS7 also.

$ telegraf --config telegraf.conf
2020-10-14T09:54:56Z I! Starting Telegraf 1.15.3
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x30 pc=0x1f94444]

goroutine 98 [running]:*Converter).convertFields(0xc0006d58f0, 0x2f59f00, 0xc00051ec40)
/go/src/ +0x6e4*Converter).Apply(0xc0006d58f0, 0xc0000a0e70, 0x1, 0x1, 0xc0000a0e50, 0xc000237cc0, 0x1)
/go/src/ +0x9c*streamingProcessor).Add(0xc00000f060, 0x2f59f00, 0xc00051ec40, 0x2f4aea0, 0xc000116040, 0x1, 0x101)
/go/src/ +0x81*RunningProcessor).Add(0xc0002006f0, 0x2f59f00, 0xc00051ec40, 0x2f4aea0, 0xc000116040, 0x0, 0x0)
/go/src/ +0x106*Agent).runProcessors.func1(0xc0006e6030, 0xc000591000)
/go/src/ +0x156
created by*Agent).runProcessors
/go/src/ +0x90

Can confirm also, the issue is not presented in 1.14.5. Any feedback would be appreciated.

This issue is fixed in 1.15.4 and 1.16.0. It’s trying to tell you “error converting to unsigned”, but there was an issue with the logger setup on our side. In 1.15.4 and above you’ll now see the error rather than a panic.