[OmniOS-discuss] High Availability and clustering on OmniOS

[Vincenzo Pii]

2014-08-28 13:26 GMT+02:00 Jim Klimov <jimklimov at cos.ru>:

> 28 августа 2014 г. 12:24:17 CEST, Theo Schlossnagle <jesus at omniti.com>
> пишет:
> >It depends on what services you are intending to make highly available.
> >We
> >do this got load balancing, http acceleration and nat using vippy. It
> >is
> >not really suited toward making services like NFS fail over.
> >On Aug 28, 2014 4:47 AM, "Vincenzo Pii" <piiv at zhaw.ch> wrote:
> >
> >> Hello,
> >>
> >> What software/technologies can be used on OmniOS to get an
> >active/passive
> >> setup between two (OmniOS) nodes?
> >>
> >> Basically, one node should be up and running all the time and, in
> >case of
> >> failures, the second one should take over transparently.
> >>
> >> Is the IHAC project (https://www.illumos.org/projects/ihac) still
> >alive
> >> or forked and maintained somewhere else? When trying to access the
> >> Repository I get a 404...
> >>
> >> Many thanks,
> >> Vincenzo.
> >>
> >> --
> >> Vincenzo Pii
> >> Researcher, InIT Cloud Computing Lab
> >> Zurich University of Applied Sciences (ZHAW)
> >> http://www.cloudcomp.ch/
> >>
> >> _______________________________________________
> >> OmniOS-discuss mailing list
> >> OmniOS-discuss at lists.omniti.com
> >> http://lists.omniti.com/mailman/listinfo/omnios-discuss
> >>
> >>
> >
> >
> >------------------------------------------------------------------------
> >
> >_______________________________________________
> >OmniOS-discuss mailing list
> >OmniOS-discuss at lists.omniti.com
> >http://lists.omniti.com/mailman/listinfo/omnios-discuss
>
> Probably, something of the kind might be done by serving the
> applications/databases/etc. from local zones, hosted along with HA data on
> storage equally available to two head-nodes.
>
> Then the failover is about grabbing the pool, and starting the zones.
> Securely mastering the zfs pool, as well as STONITH and fencing is the
> tricky part.
>
> Firing up the services (pool and zones and maybe VMs) is simple and can be
> SMFized to do in proper order as dependencies are met, I blogged on that
> with code snippets in OI/illumos wikis somewhere ;)
>
> HTH, Jim
> --
> Typos courtesy of K-9 Mail on my Samsung Android
>


How would you detect a failure in this case?
I mean, which clustering software would you use for this scenario?

Many thanks!

-- 
Vincenzo Pii
Researcher, InIT Cloud Computing Lab
Zurich University of Applied Sciences (ZHAW)
http://www.cloudcomp.ch/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://omniosce.org/ml-archive/attachments/20140828/915c73bb/attachment.html>