segfault when running

[grobian]

I think I might be doing something wrong here, but I frequently get a segfault on both linux and darwin like this:

(ptah:working-repos/carbon-c-relay) % ~/go/bin/haggar -carbon=amun:2103 -prefix=prefix1
2017/08/02 10:09:28 master: pid 10469
2017/08/02 10:09:28 agent 0: launched
2017/08/02 10:09:38 agent 0: flushed 10000 metrics
2017/08/02 10:09:39 agent 1: launched
2017/08/02 10:09:39 agent 2: launched
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x20 pc=0x4d9d16]

goroutine 18 [running]:
main.(*Agent).flush.func1(0xc420043f50)
        /home/grobian/go/src/github.com/gorsuch/haggar/main.go:60 +0x26
main.(*Agent).flush(0xc420043f50, 0x59b4a0, 0xc4200a26e0)
        /home/grobian/go/src/github.com/gorsuch/haggar/main.go:70 +0x2db
main.(*Agent).Start(0xc420043f50)
        /home/grobian/go/src/github.com/gorsuch/haggar/main.go:41 +0x71
main.launchAgent(0x0, 0x2710, 0x2540be400, 0x7fff8a1842b3, 0x9, 0x7fff8a1842c5, 0x7)
        /home/grobian/go/src/github.com/gorsuch/haggar/main.go:86 +0xeb
created by main.main
        /home/grobian/go/src/github.com/gorsuch/haggar/main.go:128 +0x3a6

Thanks for this nice tool! Any suggestions or ideas on what I'm doing wrong here?
(Or could it be the relay closing sockets on you without you expecting that?)

[gorsuch]

Hi @grobian -- thanks for submitting this issue. I was out of town and didn't get a chance to look at it until now.

I submitted a potential fix here -- #24.

If you have the cycles, I'd love for you to give things a shot now and see if the problem persists.

[grobian]

FYI I worked around the issue by using -cache_connections. Running off the fix-24 branch now, up to 3.4M metrics, which it never made before without caching connections, so it seems fixed!

[gorsuch]

fantastic -- thank you for the review and for reporting this bug! I'll merge it in.