[distcc] Problems with distcc hanging on large compiles (Patch not effective)

[Hien D. Ngo]

No joy, same hang ups, same backtrace, same log entries.  Seems to get stuck going 
to the same machine as before.  I'll remove this particular host from the mix to see 
if things improve at all.

As for my "cunning ruse", I've always had an innate backlash against the X 
convention of calling the display the server and the programs the clients (even 
though the "server" provides services such as the screen, keyboard and mouse.)  Just 
somehow seems counter intuitive... Same logic for distcc (distcc = runs on server, 
distccd = runs on clients :)

Hien

---- Original Message ----
From:		Martin Pool
Date:		Wed 8/28/02 0:39
To:		Hien D. Ngo
Cc:		distcc at lists.samba.org
Subject:	Re: [distcc] Problems with distcc hanging on large compiles (More 
debugging info)

On 27 Aug 2002, "Hien D. Ngo" <hien at moses.xp.com> wrote:
> 
> Here's what I've got.  Looks like the client completes the compile, sends the data 
> back to the server, and tries to close its connection.  The server receives all of 
> the DOTO data back, but never gets to the "completed ok" or dcc_exit stage.  The 
> server in this case is a RH 7.2 box running 2.4.9, the client is an older RH 6.x 
box 
> running 2.2.16.

Your cunning ruse of labelling the server trace as "client" was
successful :-)

Could you try this patch?

-- 
Martin 

--- serve.c.~1.31.~	2002-08-15 11:59:54.000000000 +1000
+++ serve.c	2002-08-28 16:37:29.000000000 +1000
@@ -274,6 +274,8 @@ int dcc_accept_job(int fd)
            "input file: %ld bytes; output file: %ld bytes",
            (long) i_size, (long) o_size);
 
+    tcp_cork_sock(STDIN_FILENO, 0);
+
     return ret;
 }
 
_______________________________________________
distcc mailing list
distcc at lists.samba.org
http://lists.samba.org/cgi-bin/mailman/listinfo/distcc