Lustre discuss - Nov 2007 - Problem mounting Lustre from another network

Hello,

I''ve got Lustre now set up using two networks. When the file system was
built, only one of these networks was present and configured. I have no
problems communicating with Lustre from hosts on the network that it was
built with, but when I try to mount from the network I''ve just added, I
get
this error:

-- client --
mount -t lustre tm0-0 at tcp1:tm0-1 at tcp1:/lustre /mnt/lustremount.lustre:
mount
tm0-0 at tcp1:tm0-1 at tcp1:/lustre at /mnt/lustre failed: No such file or
directory
Is the MGS specification correct?
Is the filesystem name correct?
If upgrading, is the copied client log valid? (see upgrade docs)
-- client --

I can see all of the various operational units (OSS nodes, primary and
secondary MDS nodes) both over regular TCP/IP and via the various
''ltcl''
commands (peer_list, conn_list), and can see the client there. However, if I
run ''lctl dl'' on the client, nothing shows up.

Why is this happening, and what can I do to fix it? I''ve scoured the
''net
and list archives, but I haven''t found anything that sheds any light on
the
problem.

thanks,
Klaus

On Thu, Nov 15, 2007 at 05:27:22PM -0800, Klaus Steden
wrote:
> 
> Hello,
> 
> I''ve got Lustre now set up using two networks. When the file
system was
> built, only one of these networks was present and configured. I have no
> problems communicating with Lustre from hosts on the network that it was
> built with, but when I try to mount from the network I''ve just
added, I get
> this error:
> 
> -- client --
> mount -t lustre tm0-0 at tcp1:tm0-1 at tcp1:/lustre
/mnt/lustremount.lustre: mount
> tm0-0 at tcp1:tm0-1 at tcp1:/lustre at /mnt/lustre failed: No such file or
> directory
On the client, can you please run these commands:
lctl list_nids
lctl ping tm0-0 at tcp1
lctl ping tm0-1 at tcp1

What do they say?

Isaac

Hiyo,

Here''s what I get:

-- lctl --
lctl > list_nids
172.16.128.100 at tcp1
lctl > ping tm0-0
Can''t parse process id "tm0-0"
lctl > ping tm0-1
Can''t parse process id "tm0-1"
lctl > ping 172.16.128.252 at tcp1
12345-0 at lo
12345-172.16.129.252 at tcp
12345-172.16.128.252 at tcp1
lctl > ping 172.16.128.249 at tcp1
12345-0 at lo
12345-172.16.129.249 at tcp
12345-172.16.128.249 at tcp1
-- lctl --

172.16.128.252 is the IP of tm0-0, and 172.16.128.249 is the IP of tm0-1.
Right now, they''re only listed in /etc/hosts, the client
doesn''t have useful
DNS service.

Klaus

On 11/15/07 7:06 PM, "Isaac Huang" <He.Huang at Sun.COM>did etch
on stone
tablets:
> On Thu, Nov 15, 2007 at 05:27:22PM -0800, Klaus Steden wrote:
>> 
>> Hello,
>> 
>> I''ve got Lustre now set up using two networks. When the file
system was
>> built, only one of these networks was present and configured. I have no
>> problems communicating with Lustre from hosts on the network that it
was
>> built with, but when I try to mount from the network I''ve just
added, I get
>> this error:
>> 
>> -- client --
>> mount -t lustre tm0-0 at tcp1:tm0-1 at tcp1:/lustre
/mnt/lustremount.lustre: mount
>> tm0-0 at tcp1:tm0-1 at tcp1:/lustre at /mnt/lustre failed: No such file
or
>> directory
> 
> On the client, can you please run these commands:
> lctl list_nids
> lctl ping tm0-0 at tcp1
> lctl ping tm0-1 at tcp1
> 
> What do they say?
> 
> Isaac