Re: replicating Jahia servers for redundancy purposes

Actually, you're better off to leave DNS alone altogether.  (I assume 
nobody runs Jahia without a firewall; if not, I can recommend less 
painful ways to commit suicide.)  So if one box goes down, you can just 
manually/automatically change the mapping to map both boxes to the same 
app server (IP redirection).  Or you can just use 1 IP in the 1st place, 
and let the firewall load balance.  Most modern firewalls with basic 
load balancing will not try to connect to a host that is down, and they 
will also keep track of streams, i.e. a user that is talking to one app 
server will keep talking to that one.

-- 
Cheers,
	Balázs

Serge Huber wrote:
> 
> Still on this issue, if you need to reconfigure DNS, there will be a 
> small delay for the users, because DNS propagation won't be immediate. 
> In this case why not simply restart the Jahia on server B. Surely this 
> won't take much time and you will have the system up and running. Unless 
> you were talking about automatic DNS modification, but maybe you could 
> plugin a Jahia restart in there too ?
> 
> Or maybe on server B Jahia wouldn't be running UNTIL you have a failure 
> on A, and only the filesystem updates are done. Also MySQL uses memory 
> cases, and needs to be informed that the files have changed.
> 
> cheers,
> Serge...
> 
> William Chamberlain wrote:
> 
>> Leaving aside the question of how much redundancy you want, and 
>> dealing only with what you have…
>>
>> You could run Jahia on both servers with the live filesystem and db on 
>> server A as Stéphane suggests, but keep the db and filesystem 
>> replication from A to B that you already have working. If A fails you 
>> have to reconfigure Jahia on B to use the db and filesystem on B, so 
>> it isn’t a clean switch-over but you’d always have a delay between the 
>> server dying and you realising it to switch the DNS, so you shouldn’t 
>> be much worse off. And B is always good to go as a standalone, which 
>> means you can use it as a pre-release/staging environment.
>>
>> Or you could buy ($) another few boxes and get into the whole 
>> redundancy thing and the network maintenance that goes with it.
>>
>> William Chamberlain, ex 294
>>
>> Web Master, ICTB, QSA, tel: +61 7 3864 0294
>>
>> ------------------------------------------------------------------------
>>
>> *From:* Chris Stephens [mailto:[email protected]]
>> *Sent:* Wednesday, 19 January 2005 8:28 AM
>> *To:* [email protected]
>> *Subject:* RE: replicating Jahia servers for redundancy purposes
>>
>> Hi Stéphane,
>>
>> Thanks for your comments.
>>
>> I've considered that scenario but - if I understand you correctly - it 
>> really just shares load and adds an element of redundancy. If server B 
>> goes away it doesn't matter because server A is still there. If 
>> Jahia/Tomcat fails on server A it likewise doesn't matter. But if 
>> server A dies (i.e. completely: like a ahrdware failure) you still 
>> lose the website.
>>
>> No?
>>
>> So I agree it's an improvement but it's not quite perfect.
>>
>> I was hoping there was a way to wake Jahia up to the fact that its 
>> filesystem repository had changed without restarting it. Maybe this is 
>> not possible...
>>
>> :(
>>
>> Chris
>>
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> This email (including any attached files) is for the intended 
>> recipient(s) only. If you received this email by mistake, please, as a 
>> courtesy, tell the sender, then delete this email.
>>
>> The views and opinions are the originator's and do not necessarily 
>> reflect those of the Queensland Studies Authority. All reasonable 
>> precautions have been taken to ensure that this email contained no 
>> viruses at the time it was sent.
>>  
>>
>