Allow all UDP also, LAM/MPI uses UDP for some operations and gives really really strange errors if it isn't working. Ganglia may also.
On 3/22/06, Michelle Chu <[EMAIL PROTECTED]> wrote: > > Strange to me... I disabled the SElinux and unchecked the pfilter during the > installation, but i didn't disable firewall. However, I modified the TCP > wrapper at the server node to allow all traffic from private network. Also, > when i test cluster setup, it failed at ganglia setup test, i think ganglia > is just a monitoring program and it won't affect the normal function of my > cluster. May i just ignore it? Or shall i download ganglia rpm and install > it seperatedly? > Thanks, > Michelle > > > On 3/22/06, Bernard Li <[EMAIL PROTECTED]> wrote: > > > > > > That's strange... this shouldn't happen. During installation of your > headnode, did you disable the firewall? > > > > Anyways, glad that it works now. > > > > Cheers, > > > > Bernard > > > > > > ________________________________ > > > From: Michelle Chu [mailto:[EMAIL PROTECTED] > > > > Sent: Wednesday, March 22, 2006 13:06 > > > > To: Bernard Li > > Cc: [email protected] > > Subject: Re: [Oscar-users] client installations problem > > > > > > > > > > > > I restarted my server node, and type ./install_cluster eth0 again, I then > CD boot my client node, wow,,, the client node is now installing.... > > wierd... like window system.. reboot > > > > thank you, Bernard for all your help. > > Michelle > > > > > > > > On 3/22/06, Bernard Li <[EMAIL PROTECTED]> wrote: > > > > > > > > > Does nmap localhost show that port as being opened? How about netstat? > > > > > > Cheers, > > > > > > Bernard > > > > > > > > > ________________________________ > > > > From: Michelle Chu [mailto:[EMAIL PROTECTED] > > > > > > Sent: Wednesday, March 22, 2006 12:03 > > > To: Bernard Li > > > Cc: [email protected] > > > > > > Subject: Re: [Oscar-users] client installations problem > > > > > > > > > > > > > > > > > > Yes. It is running. > > > [EMAIL PROTECTED] ~]# ps -ef |grep rsync > > > root 3525 1 0 13:11 ? 00:00:00 /usr/bin/rsync --daemon > --config > > > =/etc/systemimager/rsyncd.conf > > > root 4314 4021 0 14:00 pts/4 00:00:00 grep rsync > > > > > > Also, in my /etc/hosts.allow file > > > ALL: LOCAL > > > ALL: .cs.xxx.edu > > > > > > I belive it should allow all traffic from nodes within cs.xxx.edu > domain, right? > > > Michelle > > > > > > On 3/22/06, Bernard Li <[EMAIL PROTECTED]> wrote: > > > > > > > > > > > > Can you check to see if rsync is running on your headnode? > > > > > > > > # ps -ef | grep rsync > > > > > > > > If not, then turn it on by running: > > > > > > > > # /etc/init.d/systemimager-server-rsyncd start > > > > > > > > Cheers, > > > > > > > > Bernard > > > > > > > > > > > > ________________________________ > > > > > From: Michelle Chu [mailto:[EMAIL PROTECTED] > > > > > > > > Sent: Wednesday, March 22, 2006 11:47 > > > > To: Bernard Li > > > > Cc: Michael Edwards; > [email protected] > > > > > > > > Subject: Re: [Oscar-users] client installations problem > > > > > > > > > > > > > > > > > > > > > > > > > > > > Last login: Mon Mar 20 09:09:29 2006 > > > > [EMAIL PROTECTED] ~]# ifconfig -a > > > > eth0 Link encap:Ethernet HWaddr 00:10:18:18:31:30 > > > > inet addr: 10.1.1.1 Bcast: 10.1.1.255 Mask: 255.255.255.0 > > > > inet6 addr: fe80::210:18ff:fe18:3130/64 Scope:Link > > > > UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 > > > > RX packets:37748 errors:0 dropped:0 overruns:0 frame:0 > > > > TX packets:411437 errors:0 dropped:0 overruns:0 carrier:0 > > > > collisions:0 txqueuelen:1000 > > > > RX bytes:3854698 (3.6 MiB) TX bytes:40853456 (38.9 MiB) > > > > Interrupt:169 > > > > > > > > eth1 Link encap:Ethernet HWaddr 00:14:22:54:B3:5D > > > > inet addr:131.xxxxxx Bcast:131.xxxxxx Mask:255.255.255.0 > > > > inet6 addr: fe80::214:22ff:fe54:b35d/64 Scope:Link > > > > UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 > > > > RX packets:203143 errors:0 dropped:0 overruns:0 frame:0 > > > > TX packets:1842 errors:0 dropped:0 overruns:0 carrier:0 > > > > collisions:0 txqueuelen:1000 > > > > RX bytes:21796471 (20.7 MiB) TX bytes:233307 (227.8 KiB) > > > > Base address:0xccc0 Memory:dfce0000-dfd00000 > > > > > > > > lo Link encap:Local Loopback > > > > inet addr:127.0.0.1 Mask: 255.0.0.0 > > > > inet6 addr: ::1/128 Scope:Host > > > > UP LOOPBACK RUNNING MTU:16436 Metric:1 > > > > RX packets:693300 errors:0 dropped:0 overruns:0 frame:0 > > > > TX packets:693300 errors:0 dropped:0 overruns:0 carrier:0 > > > > collisions:0 txqueuelen:0 > > > > RX bytes:82472067 (78.6 MiB) TX bytes:82472067 ( 78.6 MiB) > > > > > > > > sit0 Link encap:IPv6-in-IPv4 > > > > NOARP MTU:1480 Metric:1 > > > > RX packets:0 errors:0 dropped:0 overruns:0 frame:0 > > > > TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 > > > > collisions:0 txqueuelen:0 > > > > RX bytes:0 (0.0 b) TX bytes:0 (0.0 b) > > > > > > > > [EMAIL PROTECTED] ~]# > > > > I have two NIC cards. One for outside internet, another for private > network. > > > > Thanks, > > > > Michelle > > > > > > > > > > > > On 3/22/06, Bernard Li <[EMAIL PROTECTED]> wrote: > > > > > > > > > > > > > > > Can you give us the output of "ifconfig" on your headnode? Please > remember to hide any sensitive IP information if necessary. I just need to > find out what your network setup is like ( e.g. how many network cards you > have, ip, etc.) > > > > > > > > > > Cheers, > > > > > > > > > > Bernard > > > > > > > > > > > > > > > ________________________________ > > > > > > From: Michelle Chu [mailto:[EMAIL PROTECTED] > > > > > > > > > > Sent: Wednesday, March 22, 2006 11:33 > > > > > To: Michael Edwards > > > > > Cc: Bernard Li; [email protected] > > > > > > > > > > Subject: Re: [Oscar-users] client installations problem > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hey, Michael, > > > > > I tried the boot CD option, and the client nodes starts to install > from the system image file, however, I am seeing another problem: > > > > > > ******************************************************************* > > > > > bla bla... > > > > > we have connectivity to your systemImager server > > > > > start_syslogd > > > > > get_boel_binaries_tarball > > > > > rsync -av 10.1.1.1: boot/../boel_binaries.tar.gz > > > > > rsync: failed to connect to 10.1.1.1 ( note: it is my server node) > > > > > connection refused. (111) > > > > > rsync: error: error in scket IO(code10) at client server > > > > > killing off running processes. > > > > > > ************************************************************************* > > > > > > > > > > How may i set in the server node in order to allow all traffic from > client nodes in the private network? > > > > > > > > > > Thanks, > > > > > Michelle > > > > > > > > > > > > > > > > > > > > On 3/21/06, Michael Edwards <[EMAIL PROTECTED] > wrote: > > > > > > Have you tried either the boot CD option or etherboot? > Occasionally I > > > > > > have had problems with TFTP not working on its own, but it works > ok > > > > > > off the boot CD. Etherboot can go on a floppy if CDs aren't an > > > > > > option, or that doesn't work either. > > > > > > > > > > > > On 3/20/06, Bernard Li < [EMAIL PROTECTED]> wrote: > > > > > > > > > > > > > > > > > > > > > Hi Michelle: > > > > > > > > > > > > > > Everything looks fine except for the following: > > > > > > > > > > > > > > Tk::Error: Failed to AUTOLOAD 'Tk::Label::selectionRange' at > > > > > > > > /usr/lib/systeminstaller/SystemInstaller/Tk/AddClients.pm > > > > > > > line 193 > > > > > > > Carp::croak at /usr/lib/perl5/5.8.5/Carp.pm line 201 > > > > > > > Tk::Widget::__ANON__ at > > > > > > > > /usr/lib64/perl5/vendor_perl/5.8.0/x86_64-linux-thread-multi/Tk/Widget.pm > > > > > > > line 347 > > > > > > > > SystemInstaller::Tk::AddClients::run_addclients at > > > > > > > > /usr/lib/systeminstaller/SystemInstaller/Tk/AddClients.pm > > > > > > > line 193 > > > > > > > Tk callback for .toplevel.button10 > > > > > > > Tk::__ANON__ at > > > > > > > > /usr/lib64/perl5/vendor_perl/5.8.0/x86_64-linux-thread-multi/Tk.pm > > > > > > > line 247 > > > > > > > Tk::Button::butUp at > > > > > > > > /usr/lib64/perl5/vendor_perl/5.8.0/x86_64-linux-thread-multi/Tk/Button.pm > > > > > > > line 111 > > > > > > > <ButtonRelease-1> > > > > > > > (command bound to event) > > > > > > > > > > > > > > Do you remember what you did when that happen? > > > > > > > > > > > > > > BTW, on your headnode, how many network cards do you have? Do > you have 2 > > > > > > > where one is on your university LAN and the other is local to > your cluster > > > > > > > nodes? > > > > > > > > > > > > > > Also, what network cards is on your compute node, they support > PXE-boot > > > > > > > right? > > > > > > > > > > > > > > P.S. When replying, please hit on "reply all" such that the > message gets > > > > > > > sent back to the oscar-users mailing-list and can be archived. > > > > > > > > > > > > > > Thanks, > > > > > > > > > > > > > > Bernard > > > > > > > > > > > > > > ________________________________ > > > > > > > > > > > > > > From: Michelle Chu [mailto: [EMAIL PROTECTED] > > > > > > > Sent: Mon 20/03/2006 20:03 > > > > > > > > > > > > > > To: Bernard Li > > > > > > > Subject: Re: [Oscar-users] client installations problem > > > > > > > > > > > > > > > > > > > > > > > > > > > > Bernard, > > > > > > > I am attaching the installation log file. I did enable dhcpd and > manully > > > > > > > assign each recoginized MAC address to each client name, such as > > > > > > > oscarnode1.domainname... > > > > > > > Thanks, > > > > > > > > > > > > > > Michelle > > > > > > > > > > > > > > On 3/20/06, Bernard Li < [EMAIL PROTECTED]> wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Instead of us trying to guess what you've done, can you > perhaps post your > > > > > > > oscarinstall.log (located in /opt/oscar) before we continue with > > > > > > > troubleshooting? > > > > > > > > > > > > > > > > A few more questions: > > > > > > > > - did you enable DHCP? > > > > > > > > - did you assign the MAC address of the node to the node that > was defined? > > > > > > > > > > > > > > > > Cheers, > > > > > > > > > > > > > > > > Bernard > > > > > > > > > > > > > > > > ________________________________ > > > > > > > From: Michelle Chu [mailto: [EMAIL PROTECTED] > > > > > > > > Sent: Mon 20/03/2006 19:02 > > > > > > > > To: Bernard Li > > > > > > > > Subject: Re: [Oscar-users] client installations problem > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Bernard, > > > > > > > > Thank you so much for you reply. Yes. I did this step before > clicking the > > > > > > > close button in setup networking. Any other suggestion? Michelle > > > > > > > > > > > > > > > > > > > > > > > > On 3/20/06, Bernard Li < [EMAIL PROTECTED]> wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > Hi Michelle: > > > > > > > > > > > > > > > > > > In the "Setup Networking" step, there is a button which > reads "Setup > > > > > > > network boot" - did you happen to click it? > > > > > > > > > > > > > > > > > > Cheers, > > > > > > > > > > > > > > > > > > Bernard > > > > > > > > > > > > > > > > > > > > > > > > > > > ________________________________ > > > > > > > From: [EMAIL PROTECTED] > [mailto: > > > > > > > [EMAIL PROTECTED] On > Behalf Of > > > > > > > Michelle Chu > > > > > > > > > Sent: Monday, March 20, 2006 16:12 > > > > > > > > > To: [email protected] > > > > > > > > > Subject: [Oscar-users] client installations problem > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hey, there, > > > > > > > > > > > > > > > > > > I encountered problem after setup networking and netboot the > client > > > > > > > nodes. Client node shows: > > > > > > > > > > > > > > > > > > client IP: 10.1.1.2 Mask: 255.255.255.0 > > > > > > > > > DHCP IP 10.1.1.1 Gateway IP: 10.1.1.1 > > > > > > > > > PXE-E32 > > > > > > > > > TFTP open timeout > > > > > > > > > TFTP... > > > > > > > > > Eventually timeout > > > > > > > > > > > > > > > > > > The tftp at server node is on by typing: chkconfig tftp on > > > > > > > > > I also disable iptables by typing: chkconfig --level=2345 > iptables off > > > > > > > > > > > > > > > > > > I guess there might be some problem with the firewall > settings. Any > > > > > > > ideas?? > > > > > > > > > > > > > > > > > > Michelle > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > ------------------------------------------------------- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642 _______________________________________________ Oscar-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/oscar-users
