You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
> Before filing a bug, please check whether it has been fixed since
> the latest release: run "hg pull -u" and retry what you did to
> reproduce the problem. Thanks.
Did this; problem still exists.
What steps will reproduce the problem?
1. Compile and run the attached go program (starts a bunch of goroutines
that try to open a bunch of TCP connections).
2.
3.
What is the expected output? What do you see instead?
There should be no output. Instead, I see the results in the attached .txt
file (somewhat truncated, but it's 11M unzipped).
What is your $GOOS? $GOARCH?
Linux, amd64
Which revision are you using? (hg identify)
88d305c3f261 tip
Please provide any additional information below.
Feel free to contact me if you need more information. I think the provided
.go source is a fairly minimal test (I *think* it only occurs from
goroutines, and only with a fairly large number of attempts to bogus,
non-localhost ports). I could be wrong, though.
This isn't a socket problem. Instead, when the go runtime is unable to create another
system thread using clone(), it panics in a rather unfriendly way. You get the same
result if you replace net.Dial() with syscall.Sleep(1e11). Note also that decreasing
the thread limit by somthing like "ulimit -u 500" will make the problem easier to
reproduce - you only need about 500 goroutines to cause the panic.
by konrad.meyer:
Attachments:
The text was updated successfully, but these errors were encountered: