This sounds suspiciously like an issue with pfilter.  You have two options:
 
1) Disable pfilter on the headnode:
 
# /etc/init.d/pfilter stop
 
then re-image your nodes
 
2) Update the kernel of the *headnode*
 
# yum update kernel-smp
 
then re-image your nodes
 
If you went with 1), note that pfilter will be restarted when you run the 
"Complete Cluster Setup" step.
 
Hopefully this works...
 
Cheers,
 
Bernard

________________________________

From: [EMAIL PROTECTED] on behalf of liukai
Sent: Mon 31/07/2006 21:58
To: [email protected]
Subject: Re: [Oscar-users] client nodes cannotreboot properlyafter imagetransfer


Dear Bernard:
       yesterday evening, I had re-imaged the client nodes. the process froze 
before transfer completed. I got a series of rsync errors:
              ............................................
              opt/lam-7.0.6/man/man3/MPI_Group_incl.3
              opt/lam-7.0.6/man/man3/MPI_Group_intersection.3
              rsync: read error: connection timed out(110)
              rsync error:error in rsync protocol data stream(code 12) at 
io.c(584)
              rsync:connection unexpectdly closed(1095846 bytes received so 
far) [generator]
              rsync error:error in rsync protocol data stream(code 12) at 
io.c(434)
              killing off running processes
              write_variables
              <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
              your autoinstall has failed.........
              ..............................
              <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
              /scripts/pre-install#_

         
      My system: FC3+ oscar4.2 , is this a systemimager problem?

      Thank you!


        Imaging logs are not stored in the headnode, you'll need to re-image 
the node and look at the console.  You can hit shift-pageup on the console to 
scroll up.
         
        Even if /mnt/sysimage did not automatically mount, you should be able 
to mount the /boot partition manually.  Assuming you used the default partition 
scheme from scsi.disk, then execute:
         
        mount /dev/sda1 /mnt
         
        Then you should be able to read /mnt/grub/menu.lst.
         
        Cheers,
         
        Bernard
        
        ________________________________
        
        From: [EMAIL PROTECTED] on behalf of liukai
        Sent: Mon 31/07/2006 00:38
        To: [email protected]
        Subject: Re: [Oscar-users] client nodes cannot reboot properlyafter 
imagetransfer
        
        
        
            thanks for your reply.
            i chose "beep" as my post-instal action , so i had to reboot clients
        manually.
            i agree with you, the transferation failed in all probability..but i
        did not yet check the error messages,  where can i find them in server 
node?
            i did try boot one of them with rescue CD, but the client's
        filesystem can't be mounted to /mnt/sysimage , so i can't see the
        context of /boot/grub.
        
            thank you again..
        
          

                When you build the client image, which post-install action did 
you choose (beep, shutdown or reboot)?  If you chose reboot it should 
automatically reboot.  Anyways, dropping to "ash" sounds like the process 
failed - did you see any other messages?
                
                Can you boot up one of the nodes with a rescue CD (you can use 
the first CD of FC3 and type "linux rescue") and then post the contents of your 
HD's /boot/grub/menu.lst?
                
                Cheers,
                
                Bernard
                
                
                ________________________________
                
                From: [EMAIL PROTECTED] on behalf of ??
                Sent: Mon 31/07/2006 00:03
                To: [email protected]
                Subject: [Oscar-users] client nodes cannot reboot properly 
after imagetransfer
                
                
                Cluster HD info:
                  > headnode   2x Piii 1.0G     35G scsi    2G ram      
                  > clientnode   2x piii 1.0G     8G  scsi     2G ram  (3 
client nodes)
                
                System:FC3 (workstation) + OSCAR4.2
                ____________________________________________
                
                     After image transfer completely (i'm not sure whether it 
is successful or not), all of the client nodes enter a "ash" , i can see the 
image filesystem in "/a" ,  then i reboot clients from hard disk manually. i 
get following errors from all of the client nodes:
                         grub loading ,step 1.5
                         please wait....
                         error 15      
                
                then I checked the manual of grub, i get these:
                        error 15 : File not found
                        This error is returned if the specified file name 
cannot be found, but everything else (like the disk/partition info) is OK.
                
                       does it imply  the transferation of image was failed? 
how can i fix it?
                  
                       Any help would be appreciated.
                
                
                ---------
                liukai
                
                 
                
------------------------------------------------------------------------
                
                
-------------------------------------------------------------------------
                Take Surveys. Earn Cash. Influence the Future of IT
                Join SourceForge.net's Techsay panel and you'll get the chance 
to share your
                opinions on IT & business topics through brief surveys -- and 
earn cash
                
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
                
------------------------------------------------------------------------
                
                _______________________________________________
                Oscar-users mailing list
                [email protected]
                https://lists.sourceforge.net/lists/listinfo/oscar-users
                 
                    

        
-------------------------------------------------------------------------
        Take Surveys. Earn Cash. Influence the Future of IT
        Join SourceForge.net's Techsay panel and you'll get the chance to share 
your
        opinions on IT & business topics through brief surveys -- and earn cash
        
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
        _______________________________________________
        Oscar-users mailing list
        [email protected]
        https://lists.sourceforge.net/lists/listinfo/oscar-users
        
        
          
        
________________________________


        
-------------------------------------------------------------------------
        Take Surveys. Earn Cash. Influence the Future of IT
        Join SourceForge.net's Techsay panel and you'll get the chance to share 
your
        opinions on IT & business topics through brief surveys -- and earn cash
        
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
        
________________________________


        _______________________________________________
        Oscar-users mailing list
        [email protected]
        https://lists.sourceforge.net/lists/listinfo/oscar-users
          


<<winmail.dat>>

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Oscar-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/oscar-users

Reply via email to