High-availability cluster and samba

[Nicolas Williams]

On Mon, 7 Feb 2000, Frederic Dubuy wrote:
> Nicolas Williams (Nicolas.Williams at wdr.com) s'exprimait ainsi : 
> 
> > Veritas HA is based on the takeover-the-failed-host's-IP-addresses-and- 
> > services model. 
> 
> It's exactly the same principle we used, the 
> takeover-the-failed-host's-IP-addresses-and-services as you call it. 
> > 
> > But note that when you take over a failed host's IP addresses you're not 
> > changing the resolution of the failed host's _names_ to a different set 
> > of IP addresses. 
> 
> Sorry if I wasn't that clear (perhaps 'cause of my approximate english). The 
> situation was we had two servers : 
> - one called cluster1, with IP 192.168.2.21 
> - one called cluster2, with IP 192.168.2.22 
> 
> As we wanted transparency for clients, we couldn't provide a service sometimes 
> called cluster1 sometimes cluster2, as it may be perturbating and prevents 
> permanent connections. That's why we used a third IP address (192.168.2.23) 
> associated with a third name : cluster3. The server running samba gets both. 
> When it fails, this IP address is used by the second server. 
> The problem is that samba tends to announce on all its interfaces (don't flame 
> me if it's not really true, I didn't look at the source code for that. It's 
> just an observation), and a windows client doesn't always see 'cluster3' as we 
> wanted to, but also cluster1 or cluster2. That's why the 'interfaces' 
> parameter is limited to the 'virtual' IP address 192.168.2.23, and the Netbios 
> name set to cluster3 (so the two smb.conf are ABSOLUTELY identical!). Using this 
>  parameters windows clients see only one server, always called cluster3. 
> The use of a Wine server and a remote announce to it shortens the latency time 
> during which some clients don't see the server any more. 

Ah, bon. J'comprend votre systeme maintenant.

Samba does honor the interfaces parameter though. Thus we don't have a
"third" IP address to which the services actually bind. Instead each
member of an HA cluster offers services at once and, if one fails the
other will take over the IP addresses and services. This aproach is a
bit more economical than the "third IP address" system because at least
you get something out of both HA cluster membes during normal
operation.

That said it's not the only valid HA configuration. It's just the one we
use. Yours is fine too.

> [...] 
> 
> Frederic 
> 

Merci,

Nico
-DISCLAIMER: an automatically appended disclaimer may follow. By posting-
-to a public e-mail mailing list I hereby grant permission to distribute-
-and copy this message.-

This message contains confidential information and is intended only 
for the individual named.  If you are not the named addressee you 
should not disseminate, distribute or copy this e-mail.  Please 
notify the sender immediately by e-mail if you have received this 
e-mail by mistake and delete this e-mail from your system.

E-mail transmission cannot be guaranteed to be secure or error-free 
as information could be intercepted, corrupted, lost, destroyed, 
arrive late or incomplete, or contain viruses.  The sender therefore 
does not accept liability for any errors or omissions in the contents 
of this message which arise as a result of e-mail transmission.  If 
verification is required please request a hard-copy version.  This 
message is provided for informational purposes and should not be 
construed as a solicitation or offer to buy or sell any securities or 
related financial instruments.