<p dir="ltr">I've seen you talking about the new clustering approach, can you elaborate a little (maybe in a separate topic/mail)?.</p>
<p dir="ltr">Given the state of distributed storage solution, shouldn't that be considered?. For example Nutanix adds capacity with each new node and the interface is just NFS. Wouldn't that make it easier for the management of upper layers?</p>
<p dir="ltr">Pomegranate or LyzardFS might be an option?</p>
<p dir="ltr">Regards,<br>
Ciro</p>
<p dir="ltr">El abr 20, 2015 6:39 PM, "Bron Gondwana" <<a href="mailto:brong@fastmail.fm">brong@fastmail.fm</a>> escribió:<br>
><br>
> I would wait for 2.5.1, which should be out in a day or so. There were<br>
> some XFER bugs in 2.5.0.<br>
><br>
> The IO hit will have to be taken regardless, it's just deferred<br>
> slightly. The 2.5 backend will work with 2.2 proxies just fine, though<br>
> of course most of the new features won't be visible to your clients,<br>
> because 2.2 gives a much reduced capability string.<br>
><br>
> Longer term, we're looking at a full unified clustering system which might<br>
> still include murder or might be totally separate. It's going to be very nice,<br>
> but it will only work for 3.0+ servers.<br>
><br>
> Bron.<br>
><br>
> On Tue, Apr 21, 2015, at 08:07 AM, Michael Sofka wrote:<br>
> > On 2015-04-20 17:16, <a href="mailto:ktm@rice.edu">ktm@rice.edu</a> wrote:<br>
> > > On Mon, Apr 20, 2015 at 05:11:00PM -0400, Michael D. Sofka wrote:<br>
> > >> Under the scenario, would 2.5 work better?<br>
> > >><br>
> > >> Mike<br>
> > >><br>
> > > Hi Mike,<br>
> > ><br>
> > > In our case, the unconstrained I/O caused by the mandatory mailbox<br>
> > > format conversion on first use would have necessitated a prolonged<br>
> > > service outage to prevent overloading the system. 2.5 will allow you<br>
> > > to schedule your conversions while the system is functional. This<br>
> > > may not be a concern for you.<br>
> ><br>
> ><br>
> > Hum, it might.... This would drive up the load on the 2.4 system as<br>
> > I'm moving mailboxes?<br>
> ><br>
> > This project is driven entirely by the state of the SAN disks. They<br>
> > are either old with controller errors, or expensive to keep on<br>
> > service, or needed elsewhere in a chain of updates. Plan B is to<br>
> > clone the existing<br>
> > 2.3 server, but if I can get a new OS and application image in the<br>
> > process, I will be a happy camper. But even doing that is exceeding<br>
> > my mandate.<br>
> ><br>
> > But if a 2.5 image will work with 2.2 front-end proxies, the deferred<br>
> > conversion is worth considering. I do anticipate the moves being off-<br>
> > hours, but even off-hours is busy.<br>
> ><br>
> > Mike<br>
> ><br>
> ><br>
> > --<br>
> > Michael D. Sofka <a href="mailto:sofkam@rpi.edu">sofkam@rpi.edu</a> C&MT Sr. Systems<br>
> > Programmer, Email, TeX, Epistemology Rensselaer Polytechnic<br>
> > Institute, Troy, NY. <a href="http://www.rpi.edu/~sofkam/">http://www.rpi.edu/~sofkam/</a><br>
> ><br>
> > ----<br>
> > Cyrus Home Page: <a href="http://www.cyrusimap.org/">http://www.cyrusimap.org/</a> List Archives/Info:<br>
> > <a href="http://lists.andrew.cmu.edu/pipermail/info-cyrus/">http://lists.andrew.cmu.edu/pipermail/info-cyrus/</a> To Unsubscribe:<br>
> > <a href="https://lists.andrew.cmu.edu/mailman/listinfo/info-cyrus">https://lists.andrew.cmu.edu/mailman/listinfo/info-cyrus</a><br>
><br>
><br>
> --<br>
> Bron Gondwana<br>
> <a href="mailto:brong@fastmail.fm">brong@fastmail.fm</a><br>
> ----<br>
> Cyrus Home Page: <a href="http://www.cyrusimap.org/">http://www.cyrusimap.org/</a><br>
> List Archives/Info: <a href="http://lists.andrew.cmu.edu/pipermail/info-cyrus/">http://lists.andrew.cmu.edu/pipermail/info-cyrus/</a><br>
> To Unsubscribe:<br>
> <a href="https://lists.andrew.cmu.edu/mailman/listinfo/info-cyrus">https://lists.andrew.cmu.edu/mailman/listinfo/info-cyrus</a><br>
</p>