Do you have a nid on the MDS for the NAT'ed Network?  The mds should
have nids  [EMAIL PROTECTED] and say [EMAIL PROTECTED] (I made that up),
then you could connect to the MDS on the other network(NAT'ed) interface

Evan

> -----Original Message-----
> From: [EMAIL PROTECTED] 
> [mailto:[EMAIL PROTECTED] On Behalf Of 
> Jim McCusker
> Sent: Monday, June 11, 2007 9:32 AM
> To: [email protected]
> Subject: [Lustre-discuss] Problem mounting volumes inside 
> vmware NAT (1.4.10)
> 
> I have a vmware server running several vms that are stored on 
> a lustre volume. When I use bridged networking, I am able to 
> mount the volume inside the vm with no problem. I would like 
> to have some vms on a private NAT'ed network as database 
> servers (with the data files stored on lustre), but when I 
> change the NIC to be on the NAT network, the mount fails with 
> the following:
> 
> [EMAIL PROTECTED] ~]# /etc/init.d/lustrefs  start Mounting 
> Lustre filesystems:  mount.lustre: 
> mount(chai.med.yale.edu:mds1/client, /vol) failed: Input/output error
> mds nid 0:       [EMAIL PROTECTED]
> mds name:        mds1
> profile:         client
> options:         rw,flock
> retry:           0
>                                                            [FAILED]
> 
> /var/log/messages shows:
> 
> Jun 11 03:13:06 localhost kernel: LustreError: 
> 2525:0:(socklnd_cb.c:2160:ksocknal_recv_hello()) Error -104 
> reading HELLO from 128.36.115.13 Jun 11 03:13:06 localhost 
> kernel: LustreError: Connection to [EMAIL PROTECTED] at host 
> 128.36.115.13 on port 988 was reset: is it running a 
> compatible version of Lustre and is [EMAIL PROTECTED] one of its NIDs?
> Jun 11 03:13:11 localhost kernel: LustreError: 
> 3931:0:(client.c:947:ptlrpc_expire_one_request()) @@@ timeout 
> (sent at 1181545986, 5s ago)  [EMAIL PROTECTED] x9/t0 
> o38->[EMAIL PROTECTED]@tcp:12 lens 240/272 ref 1 fl Rpc:/0/0 rc 0/0
> Jun 11 03:13:11 localhost kernel: LustreError: mdc_dev: The 
> configuration 'client' could not be read from the MDS 'mds1'. 
>  This may be the result of communication errors between the 
> client and the MDS, or if the MDS is not running.
> Jun 11 03:13:11 localhost kernel: LustreError: 
> 3928:0:(llite_lib.c:962:lustre_fill_super()) Unable to 
> process log: client Jun 11 03:13:11 localhost mount: mount.lustre: 
> mount(chai.med.yale.edu:mds1/client, /vol) failed: Input/output error
> Jun 11 03:13:11 localhost mount: mds nid 0:       [EMAIL PROTECTED]
> Jun 11 03:13:11 localhost mount: mds name:        mds1
> Jun 11 03:13:11 localhost mount: profile:         client
> Jun 11 03:13:11 localhost mount: options:         rw,flock
> Jun 11 03:13:11 localhost mount: retry:           0
> Jun 11 03:13:11 localhost lustrefs: Mounting Lustre 
> filesystems:  failed
> 
> Only the vmware server has more than one NIC enabled, and 
> that isn't having any trouble connecting, so "options lnet 
> networks=tcp(eth0)" 
> doesn't seem like the right option.
> 
> Jim
> 
> _______________________________________________
> Lustre-discuss mailing list
> [email protected]
> https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
> 

_______________________________________________
Lustre-discuss mailing list
[email protected]
https://mail.clusterfs.com/mailman/listinfo/lustre-discuss

Reply via email to