[OpenIndiana-discuss] OpenIndiana/ZFS mail server HA, config

[Hans J. Albertsson]

Actually, automatic failover will not be necessary, it is after all only 
a mail server in an environment with fairly resilient users.
They'd be quite happy to have to call a sysadm up for manual failover.

I think in an almost unsupported OS like OI is today, one shouldn't rely 
on anything but the most basic stuff. I wouldn't expect to ever be able 
to test a hacked-up auto failover mechanism well enough to hand it over 
to an unsuspecting colleague with little experience outside Windows or Mac.

So: manual failover is absolutely OK.


On 2011-12-06 03:39, openindiana-discuss-request at openindiana.org wrote:
> Message: 2 Date: Mon, 5 Dec 2011 17:06:33 -0600 From: Lucas Van Tol 
> <catseyev9 at hotmail.com> To: <openindiana-discuss at openindiana.org> 
> Subject: Re: [OpenIndiana-discuss] OpenIndiana/ZFS mail server HA 
> config Message-ID: <BLU151-W553A40F2CAD86D380346F5FCB50 at phx.gbl> 
> Content-Type: text/plain; charset="iso-8859-1" I had been trying to 
> work out some kind of HA using iscsi LUNS + ZFS mirror on top of that; 
> but I never got around to figuring out an automated failover process. 
> You would want the master to not share out it's LUN; and the other 
> node to check if there is a system at the 'master' address before 
> trying to become the master. It probably wouldn't be very fast, but it 
> might be sufficient if you can get the failover worked out.
>> >  Date: Tue, 6 Dec 2011 10:01:34 +1100
>> >  From:josh at sysmgr.org
>> >  To:openindiana-discuss at openindiana.org
>> >  Subject: Re: [OpenIndiana-discuss] OpenIndiana/ZFS mail server HA config
>> >  
>> >  On 6 December 2011 09:05, Hans J. Albertsson
>> >  <hans.j.albertsson at branneriet.se>  wrote:
>>> >  >  I noticed AVS: This is too recent for me to have had any chance of learning
>>> >  >  about it while at Sun.
>>> >  >  Could this be used for remote mirroring between two standard Supermicro
>>> >  >  X8STi-LN4 based microserver
>>> >  >  using two ordinary SATA disks each?
>> >  
>> >  Just spitballing, but perhaps you could do something with ZFS
>> >  send/receive.  You could take frequent incremental snapshots and send
>> >  them to the remote node.  If you did this, say, every ~30-60 seconds
>> >  during periods of high churn then perhaps all of the requisite blocks
>> >  would still be in the ARC and you could send them without reading back
>> >  from disk.
>> >  
>> >  We used to do this every 15 minutes on a pair of thumpers at the Uni I
>> >  worked at.  Has anybody done this with a shorter period?  You may get
>> >  more traction with the question on a ZFS mailing list.
>> >  
>> >  -- 
>> >  Joshua M. Clulow
>> >  UNIX Admin/Developer
>> >  http://blog.sysmgr.org
>> >  
>> >  _______________________________________________
>> >  OpenIndiana-discuss mailing list
>> >  OpenIndiana-discuss at openindiana.org
>> >  http://openindiana.org/mailman/listinfo/openindiana-discuss
>   		 	   		
>