pcp
[Top] [All Lists]

Re: [pcp] qa/861 hanging

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Subject: Re: [pcp] qa/861 hanging
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Tue, 2 Sep 2014 00:30:07 -0400 (EDT)
Cc: PCP <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <54054425.9000804@xxxxxxxxxxxxxxxx>
References: <540533AF.3030308@xxxxxxxxxxxxxxxx> <1946971949.42117047.1409627496431.JavaMail.zimbra@xxxxxxxxxx> <54054425.9000804@xxxxxxxxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: V9meaoCu0qgL6ujfnRriAKN3snP9dw==
Thread-topic: qa/861 hanging
Hi Ken,

----- Original Message -----
> Here you go ...
> 
> kenj     30526  9655  0 10:15 pts/0    00:00:00 sh 861
> kenj     30540 30526  0 10:15 pts/0    00:00:00 pmcd -H nosuchhost.com -f -x
> 861.full -l /tmp/861-30526.log

> kenj@vm20:~/src/pcp/qa$ pstree 30526
> shâââpmcd

> kenj@vm20:~/src/pcp/qa$ netstat -tnl
> Active Internet connections (only servers)
> Proto Recv-Q Send-Q Local Address           Foreign Address         State
> tcp        0      0 0.0.0.0:9876            0.0.0.0:*               LISTEN
> tcp6       0      0 :::9876                 :::*                    LISTEN

OK, so it looks like everything has gone A-OK - pmcd was started, is running,
the pminfo is probably done by this point and I'd guess we're sitting in the
shell _cleanup() routine at the call to "wait".

If so, the question is why has pmcd not terminated after we sent it sigterm?
(and why wait for it too, I suppose?...)

$ diff --git a/qa/861 b/qa/861
index bd4bdb1..1b4e747 100755
--- a/qa/861
+++ b/qa/861
@@ -27,7 +27,6 @@ _cleanup()
     then
         echo "Terminating local pmcd process"
         $signal -s TERM $pid
-        wait
         pid=""
     fi
     cd $here; rm -rf $tmp.*


But, can you connect to the pmcd PID 30540 with gdb and get a stacktrace?
I'd be interested to know what its up to, it should have exited.

cheers.

--
Nathan

<Prev in Thread] Current Thread [Next in Thread>