Hi Justin,<div><br></div><div>thanks for the fast response and the great work. I thought about taking a crack at a fix myself, but Im not familiar with the codebase and don't have very little time right now.</div><div>
<br></div><div>thanks</div><div>Paul<br><br><div class="gmail_quote">On Wed, Jan 18, 2012 at 21:33, Justin Riley <span dir="ltr"><<a href="mailto:jtriley@mit.edu">jtriley@mit.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
-----BEGIN PGP SIGNED MESSAGE-----<br>
Hash: SHA1<br>
<br>
Hi Paul,<br>
<br>
I just tested for myself and I can confirm that /etc/exports is indeed<br>
being clobbered when running the 'addnode' command. I'm working on a<br>
patch release to fix this and other minor things. Should be out tomorrow.<br>
<br>
Thanks for reporting!<br>
<br>
~Justin<br>
<div><div class="h5"><br>
On 01/18/2012 02:08 PM, Paul Koerbitz wrote:<br>
> Dear starcluster team,<br>
><br>
> I tripped over what might be an error with the NFS setup when<br>
> adding new nodes to a cluster.<br>
><br>
> I set up my cluster with initially one root node only and then<br>
> first added one node and subsequently 4 more nodes. I noticed that<br>
> my ebsvolume wasn't getting mounted correctly on the nodes, calling<br>
> 'df' reported 'stale filehandle' for /home /opt/sge6 and /data<br>
><br>
> My impression is that as nodes get added, the /etc/exports file<br>
> which is responsible for allowing NFS access gets overwritten.<br>
> Therefore only the last added node can access the shared file<br>
> systems.<br>
><br>
> Here is how I resloved the issue. First I unmounted all the<br>
> volumes:<br>
><br>
> root@node001:~# umount -f /data<br>
><br>
> At this point remounting doesn't work:<br>
><br>
> root@node001:~# mount -t nfs master:/data /data<br>
><br>
> mount.nfs: access denied by server while mounting master:/data<br>
><br>
><br>
> I then edited /etc/exports on the master node. Here only the last<br>
> node was listed:<br>
><br>
> /home node005(async,no_root_squash,no_subtree_check,rw) /opt/sge6<br>
> node005(async,no_root_squash,no_subtree_check,rw) /data<br>
> node005(async,no_root_squash,no_subtree_check,rw)<br>
><br>
> I changed this to /home<br>
> node001(async,no_root_squash,no_subtree_check,rw) /opt/sge6<br>
> node001(async,no_root_squash,no_subtree_check,rw) /data<br>
> node001(async,no_root_squash,no_subtree_check,rw) /home<br>
> node002(async,no_root_squash,no_subtree_check,rw) /opt/sge6<br>
> node002(async,no_root_squash,no_subtree_check,rw) /data<br>
> node002(async,no_root_squash,no_subtree_check,rw) /home<br>
> node003(async,no_root_squash,no_subtree_check,rw) /opt/sge6<br>
> node003(async,no_root_squash,no_subtree_check,rw) /data<br>
> node003(async,no_root_squash,no_subtree_check,rw) /home<br>
> node004(async,no_root_squash,no_subtree_check,rw) /opt/sge6<br>
> node004(async,no_root_squash,no_subtree_check,rw) /data<br>
> node004(async,no_root_squash,no_subtree_check,rw) /home<br>
> node005(async,no_root_squash,no_subtree_check,rw) /opt/sge6<br>
> node005(async,no_root_squash,no_subtree_check,rw) /data<br>
> node005(async,no_root_squash,no_subtree_check,rw)<br>
><br>
> then restart the nfs-server:<br>
><br>
> $ /etc/init.d/nfs-kernel-server restart<br>
><br>
> After that running 'df' on each node showed the nfs now working<br>
> correctly.<br>
><br>
> kind regards Paul<br>
><br>
><br>
</div></div>> This body part will be downloaded on demand.<br>
<br>
-----BEGIN PGP SIGNATURE-----<br>
Version: GnuPG v2.0.17 (GNU/Linux)<br>
Comment: Using GnuPG with Mozilla - <a href="http://enigmail.mozdev.org/" target="_blank">http://enigmail.mozdev.org/</a><br>
<br>
iEYEARECAAYFAk8XLJUACgkQ4llAkMfDcrkc3wCgi+vGwbv7fJDYmf3UBLuJp9QP<br>
06MAn2QNOt+EFuTwnaiCyemhttM6oTdo<br>
=jNrz<br>
-----END PGP SIGNATURE-----<br>
</blockquote></div><br></div>