Nfs Failover

Nfs Failover

[Saurabh Sehgal]

[-- Attachment #1.1: Type: text/plain, Size: 942 bytes --]

Hi all,

I am trying to do an nfs failover from one server to another in a 4 node
cluster . During this failover, I have a link from /var/lib/nfs directory to
a mount that resides on a shared disk. When i want to move the "resources"
from one node to the other ... this is how I go about initiating the move :

1) /etc/init.d/nfsserver stop
2) mount /var/lib/nfs on the failover node
3) move IP address to failover node
3) /etc/init.d/nfsserver start
4) /sbin/sm-notify -v <failover ip address>

Are these the correct steps ? What I end up getting is a Stale NFS File
handle on all client nodes, until I unmount and remount them again. For the
purpose of my application this mount/unmount of clients is undesirable .
What should I do ? Any help would be appreciated.

( I am using SLES 10 SP 1 as a platform :
uname -a
Linux mach1 2.6.16.46-0.12-smp #1 SMP Thu May 17 14:00:09 UTC 2007 x86_64
x86_64 x86_64 GNU/Linux  )


Thank you,

Saurabh

[-- Attachment #1.2: Type: text/html, Size: 1028 bytes --]

[-- Attachment #2: Type: text/plain, Size: 314 bytes --]

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[-- Attachment #3: Type: text/plain, Size: 140 bytes --]

_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: Nfs Failover

[J. Bruce Fields]

On Tue, Oct 09, 2007 at 03:40:22PM -0400, Saurabh Sehgal wrote:
> Hi all,
> 
> I am trying to do an nfs failover from one server to another in a 4 node
> cluster . During this failover, I have a link from /var/lib/nfs directory to
> a mount that resides on a shared disk. When i want to move the "resources"
> from one node to the other ... this is how I go about initiating the move :
> 
> 1) /etc/init.d/nfsserver stop
> 2) mount /var/lib/nfs on the failover node
> 3) move IP address to failover node
> 3) /etc/init.d/nfsserver start
> 4) /sbin/sm-notify -v <failover ip address>
> 
> Are these the correct steps ? What I end up getting is a Stale NFS File
> handle on all client nodes, until I unmount and remount them again. For the
> purpose of my application this mount/unmount of clients is undesirable .
> What should I do ? Any help would be appreciated.

What filesystem are you exporting?

--b.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: Nfs Failover

[Jordi Prats]

To avoid the stale state on your clients you should add a tag to your 
exported filesystems. For example:

/usr 10.0.0.1(rw,sync,no_root_squash,nohide,fsid=1)
/usr 10.0.0.2(rw,sync,no_root_squash,nohide,fsid=1)

/var 10.0.0.1(rw,sync,no_root_squash,nohide,fsid=2)
/var 10.0.0.2(rw,sync,no_root_squash,nohide,fsid=2)

regards,
Jordi

J. Bruce Fields wrote:
> On Tue, Oct 09, 2007 at 03:40:22PM -0400, Saurabh Sehgal wrote:
>> Hi all,
>>
>> I am trying to do an nfs failover from one server to another in a 4 node
>> cluster . During this failover, I have a link from /var/lib/nfs directory to
>> a mount that resides on a shared disk. When i want to move the "resources"
>> from one node to the other ... this is how I go about initiating the move :
>>
>> 1) /etc/init.d/nfsserver stop
>> 2) mount /var/lib/nfs on the failover node
>> 3) move IP address to failover node
>> 3) /etc/init.d/nfsserver start
>> 4) /sbin/sm-notify -v <failover ip address>
>>
>> Are these the correct steps ? What I end up getting is a Stale NFS File
>> handle on all client nodes, until I unmount and remount them again. For the
>> purpose of my application this mount/unmount of clients is undesirable .
>> What should I do ? Any help would be appreciated.
> 
> What filesystem are you exporting?
> 
> --b.
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems?  Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >> http://get.splunk.com/
> _______________________________________________
> NFS maillist  -  NFS@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/nfs
> 
> 

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: Nfs Failover

[Saurabh Sehgal]

[-- Attachment #1.1: Type: text/plain, Size: 2075 bytes --]

Hi,

I will be sure to try that and see if I get the same problem ...

I am exporting an ext3 filesystem ... Does that change anything ?

Thank you,

Saurabh

On 10/9/07, Jordi Prats <jprats@cesca.es> wrote:
>
> To avoid the stale state on your clients you should add a tag to your
> exported filesystems. For example:
>
> /usr 10.0.0.1(rw,sync,no_root_squash,nohide,fsid=1)
> /usr 10.0.0.2(rw,sync,no_root_squash,nohide,fsid=1)
>
> /var 10.0.0.1(rw,sync,no_root_squash,nohide,fsid=2)
> /var 10.0.0.2(rw,sync,no_root_squash,nohide,fsid=2)
>
> regards,
> Jordi
>
> J. Bruce Fields wrote:
> > On Tue, Oct 09, 2007 at 03:40:22PM -0400, Saurabh Sehgal wrote:
> >> Hi all,
> >>
> >> I am trying to do an nfs failover from one server to another in a 4
> node
> >> cluster . During this failover, I have a link from /var/lib/nfs
> directory to
> >> a mount that resides on a shared disk. When i want to move the
> "resources"
> >> from one node to the other ... this is how I go about initiating the
> move :
> >>
> >> 1) /etc/init.d/nfsserver stop
> >> 2) mount /var/lib/nfs on the failover node
> >> 3) move IP address to failover node
> >> 3) /etc/init.d/nfsserver start
> >> 4) /sbin/sm-notify -v <failover ip address>
> >>
> >> Are these the correct steps ? What I end up getting is a Stale NFS File
> >> handle on all client nodes, until I unmount and remount them again. For
> the
> >> purpose of my application this mount/unmount of clients is undesirable
> .
> >> What should I do ? Any help would be appreciated.
> >
> > What filesystem are you exporting?
> >
> > --b.
> >
> >
> -------------------------------------------------------------------------
> > This SF.net email is sponsored by: Splunk Inc.
> > Still grepping through log files to find problems?  Stop.
> > Now Search log events and configuration files using AJAX and a browser.
> > Download your FREE copy of Splunk now >> http://get.splunk.com/
> > _______________________________________________
> > NFS maillist  -  NFS@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/nfs
> >
> >
>

[-- Attachment #1.2: Type: text/html, Size: 2763 bytes --]

[-- Attachment #2: Type: text/plain, Size: 314 bytes --]

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[-- Attachment #3: Type: text/plain, Size: 140 bytes --]

_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

Re: Nfs Failover

[J. Bruce Fields]

On Wed, Oct 10, 2007 at 09:38:18AM -0400, Saurabh Sehgal wrote:
> I will be sure to try that and see if I get the same problem ...
> 
> I am exporting an ext3 filesystem ...

So you're using a shared block device?

> Does that change anything ?

It's just important to make sure you're exporting the *identical* ext3
filesystem from the two servers--copies made above the filesystem level
(using cp or rsync or whatever) aren't sufficient to ensure that
filehandles stay the same.

--b.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs

nfs failover

[Eric Ford]

[-- Attachment #1: Type: text/plain, Size: 742 bytes --]

Hello, everybody
 
I installed nfs on my cluster system, two head nodes(one is primary and another is backup) with two clients. They use NFS to store and transfer data between server and clients. But whenever failover happens, the client can't connect to backup server although 
backup server has the same IP and hostname and same configuration at that time.
 
"NFS server is not responding, still trying" 
 
I use this command: #service nfs restart    on backup server, it is not useful.
 
What should I do on backup server so that client can get connection with it? By the way I use Pfilter (iptables) on both servers.

Thank you

Eric


---------------------------------
Do you Yahoo!?
Yahoo! Hotjobs: Enter the "Signing Bonus" Sweepstakes

[-- Attachment #2: Type: text/html, Size: 1036 bytes --]

Re: nfs failover

[Brian Tinsley]

[-- Attachment #1: Type: text/plain, Size: 1297 bytes --]

There are some specific events that must occur in order to make NFS
server failover happen. You should search the mailing list archives for
the linux-ha software (a.k.a., heartbeat) for further details. NFS
server failover has been detailed there quite extensively over the past
couple of years. And if you do not use linux-ha for clustering
(especially in the context you describe here), I suggest you give it a
try.


On Sat, 2004-01-10 at 21:02, Eric Ford wrote:

> Hello, everybody
>  
> I installed nfs on my cluster system, two head nodes(one is primary
> and another is backup) with two clients. They use NFS to store and
> transfer data between server and clients. But whenever failover
> happens, the client can't connect to backup server although 
> backup server has the same IP and hostname and same configuration at
> that time.
>  
> "NFS server is not responding, still trying" 
>  
> I use this command: #service nfs restart    on backup server, it is
> not useful.
>  
> What should I do on backup server so that client can get connection
> with it? By the way I use Pfilter (iptables) on both servers.
> 
> Thank you
> 
> Eric
> 
> 
> ______________________________________________________________________
> Do you Yahoo!?
> Yahoo! Hotjobs: Enter the "Signing Bonus" Sweepstakes

[-- Attachment #2: Type: text/html, Size: 1851 bytes --]