Package: icecc
Version: 0.9.1-1
Severity: normal
Tags: patch
The problem is in the libicecc.a: the function
DiscoverSched::try_get_scheduler always return NULL because connect
always returns EINPROGRESS (the socket is non-blocking).
With the attached patch, we use a blocking sockect to connect to
scheduler. Tested on a 25 nodes cluster with no problems.
This bug is present in icecc-monitor=1.1-3 (compiled against icecc 0.9)
but was not present in icecc-monitor=1.1-2 (compiled against icecc 0.8).
-- System Information:
Debian Release: lenny/sid
APT prefers unstable
APT policy: (990, 'unstable'), (550, 'testing'), (25, 'experimental')
Architecture: amd64 (x86_64)
Kernel: Linux 2.6.27 (SMP w/2 CPU cores)
Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Versions of packages icecc depends on:
ii adduser 3.110 add and remove users and groups
ii debconf [debconf-2.0] 1.5.24 Debian configuration management sy
ii dpkg 1.14.22 Debian package management system
ii g++ [c++-compiler] 4:4.3.2-2 The GNU C++ compiler
ii g++-4.3 [c++-compiler] 4.3.2-2~exp1 The GNU C++ compiler
ii gcc [c-compiler] 4:4.3.2-2 The GNU C compiler
ii gcc-4.3 [c-compiler] 4.3.2-2~exp1 The GNU C compiler
ii libc6 2.8+20080809-2 GNU C Library: Shared libraries
ii libgcc1 1:4.3.2-2~exp1 GCC support library
ii libstdc++6 4.3.2-2~exp1 The GNU Standard C++ Library v3
ii lsb-base 3.2-20 Linux Standard Base 3.2 init scrip
icecc recommends no packages.
Versions of packages icecc suggests:
ii icecc-monitor 1.1-3 icecc monitor for KDE
-- debconf information excluded
*** 12_fix_nonblocking_sockets_for_icemon.diff
--- icecc-0.9.1.orig/services/comm.cpp 2008-05-12 19:57:14.000000000 +0200
+++ icecc-0.9.1/services/comm.cpp 2008-10-20 15:43:23.000000000 +0200
@@ -1160,14 +1160,21 @@
if (ask_fd >= 0)
{
+ fcntl(ask_fd, F_SETFL, 0);
int status = connect (ask_fd, (struct sockaddr*) &remote_addr,
sizeof(remote_addr) );
- if (status == 0 || (status < 0 && errno == EISCONN))
+ int connect_errno = errno;
+ fcntl(ask_fd, F_SETFL, O_NONBLOCK);
+ if (status == 0 || (status < 0 && connect_errno == EISCONN))
{
int fd = ask_fd;
ask_fd = -1;
return Service::createChannel(fd,
(struct sockaddr*) &remote_addr, sizeof(remote_addr));
}
+ else
+ {
+ log_error() << "connect failed : " << strerror(connect_errno) <<
endl;
+ }
}
return 0;
}
--- icecc-0.9.1.orig/services/comm.cpp 2008-05-12 19:57:14.000000000 +0200
+++ icecc-0.9.1/services/comm.cpp 2008-10-20 15:43:23.000000000 +0200
@@ -1160,14 +1160,21 @@
if (ask_fd >= 0)
{
+ fcntl(ask_fd, F_SETFL, 0);
int status = connect (ask_fd, (struct sockaddr*) &remote_addr,
sizeof(remote_addr) );
- if (status == 0 || (status < 0 && errno == EISCONN))
+ int connect_errno = errno;
+ fcntl(ask_fd, F_SETFL, O_NONBLOCK);
+ if (status == 0 || (status < 0 && connect_errno == EISCONN))
{
int fd = ask_fd;
ask_fd = -1;
return Service::createChannel(fd,
(struct sockaddr*) &remote_addr, sizeof(remote_addr));
}
+ else
+ {
+ log_error() << "connect failed : " << strerror(connect_errno) <<
endl;
+ }
}
return 0;
}
