Screen dump analysis:

Ao Jiang ajiang at
Tue Apr 30 20:19:05 PDT 2002

   I have some questions, when I install Scyld Beowulf (version 2.0 preview).
 I am looking forward to seeing someone can give me some direction. Thanks a

 After boot up the beowulf sys, the slave node shows:
 "Boot: System boot phase 1 in progress...
 Sending RARP request..."
 But this signal seems to send forever, status of node are always 'down'
in the Beosetup of the master node.

 I checked the website and found it may be caused by the beoserv which isn't
seeing RARP signals. But unfortunatly I couldn't find an effective way to
 solve it.

 So would you mind giving me some suggestion on how to fix beoserv problem?
 Or how to send message to slave nodes manually?

 The interesting thing is:

 When I tried to reboot the master node, the slave node seems to receive the
 messages from the master node and reboot too and enter phase 2 or 3. But
 some slave nodes show something wrong. I don't know what they mean and what the
 reasons are? If it is the problem of hardware, which device it is?
 The following is the screen dump:

 Slave node 1:

 EXT2-fs error (device ramdisk(1,3))
 Ext2_add_entry: bad entry in directory #8193; directory entry across
 blocks-offset=13680, inode=9096, rec_len=2064, name_len=5.
 The status of node 1 is error.

 Slave node 2:

 Boot: System boot phase 2 in progress...
 VFS: Cannot open root device 03:01
 Kernel panic: VFS unable to mount root fs on 03:01
 The status of node 2 is error.


More information about the Beowulf mailing list