MPI not running with Scyld

David Leunen leunen.d at fsagx.ac.be
Tue Mar 26 00:41:16 PST 2002


Hello everyone,

I installed Scyld Beowulf 27z-8.
When I boot a slave node, here are the lines that seems errors. I can 
provide the complete log file if you want. (does this list allow 
attachments?)

[...]
eth0: 3Com 3C905C Tornado  *** INVALID CHECKSUM 002F ***
[...]
hda: no response (satus = 0xa1), resetting drive
hda: no response (satus = 0xa1)
hdb: no response (satus = 0xa1), resetting drive
hdb: no response (satus = 0xa1)
[...]

It obviously looks for hard drives though we've diskless nodes ...
Here is what the /etc/beowulf/fstab looks like:

/dev/ram3      /         ext2    fs_size=65536      0 0
none           /proc     proc    defaults           0 0
none           /dev/pts  devpts  gid=5,mode=620     0 0
$MASTER:/home  /home     nfs     defaults           0 0


Even displaying these errors, the nodes complete booting sequences and 
become UP. But if I try to run an MPI program (one of the examples)
It says, on master node:

p0_7699: p4_error: net_create_slave: bproc_rfork: -1
     p4_error: latest msg from perror: Operation not permitted
bm_list_7650: p4_error: interrupt SIGINT: 2

and on the slave nodes:

vmadump: mmap failed: /usr/lib/mlibbproc.so.1.0.0

Can anyone tell me what is wrong.

Thank you.


David




More information about the Beowulf mailing list