golang goroutine practice, function or channel? - go

Question

constantly receive Json data from websocket and process them in goroutine, no idea is this writing pattern is encourage or not
ws.onmessage { //infinite receive message from websocket
go func() { //work find using this goroutine
defer processJson(message)
}()
go processJson(message) //error and program will terminated
}
func processJson(msg string) {
//code for process json
insertDatabase(processedMsg)
}
func insertDatabase(processedMsg string) {
//code insert to database
}
Below(the first goroutine) work just fine, but sometime(a week) indicates there is a data race in the code and terminate the program.
go func() {
defer processJson(message)
}()
the second goroutine, often encounter error after few minutes running, the error often is "fatal error: unexpected signal during runtime execution".
go processJson(message)
from my understanding both goroutine do the samething, why is that the first can run well and second cannot. i have try using channel, but not much difference compare to the first goroutine.
msgChan := make(chan string, 1000)
go processJson(msgChan)
for { //receive json from websocket, send to channel
msgChan <- message
}
func JsonProcessor(msg chan string) {
for { //get data from channel, process in goroutine function
msgModified := <-msg
insertDatabase(msgModified)
}
}
is there any encourage way to acheive the goal without data race, suggestions are welcome.
Appreciate and Thanks.

Answer 1

try to use sync.Mutex avoid data racing
mutux := sync.Mutex{}
ws.onmessage {
processJson(message)
}
func processJson(msg string) {
mutux.Lock()
// .........
mutux.Unlock()
}
if the processing function can be divided without data racing, multithread version as follows :
msgChan1 := make(chan string, 1000)
msgChan2 := make(chan string, 1000)
go func() {
for m := range msgChan1 {
// ...
}
}()
go func() {
for m := range msgChan2 {
// ...
}
}()
ws.onmessage {
msgChan1 <- message
msgChan2 <- message
}
ws.onclose {
close(msgChan1)
close(msgChan2)
}