<div dir="ltr">I believe Terracotta's "next big thing" is essentially a move to the Terracotta-based EHCache, which we already make use of in our solution (for other clustering needs). I agree that the Terracotta DSO mode was fairly intrusive during maintenance activities, often requiring a full restart & clean of the solution after a web application re-deploy... I am hoping that the non-DSO mode will be a bit more flexible in that regard, but only time will tell. And it probably won't solve the maintenance issues you have identified.<div>
<br></div><div>Thanks a lot for your time on this. <br><div><br></div><div style>Marc</div></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Tue, May 28, 2013 at 9:50 AM, Cantor, Scott <span dir="ltr"><<a href="mailto:cantor.2@osu.edu" target="_blank">cantor.2@osu.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">On 5/28/13 9:13 AM, "Marc Thornton" <<a href="mailto:marc.thornton@gmail.com">marc.thornton@gmail.com</a>> wrote:<br>
<br>
>I would be interested in having access to Paul's Infinispan plugin and/or<br>
>any other clustering-related plugin sources.<br>
<br>
</div><a href="https://wiki.shibboleth.net/confluence/display/SHIB2/Contributions" target="_blank">https://wiki.shibboleth.net/confluence/display/SHIB2/Contributions</a><br>
<div class="im"><br>
> My understanding with Terracotta was that there really wasn't any<br>
>Terracotta-specific code in the Shibboleth IDP, therefore any work that<br>
>may have been done to identify the integration points for an alternate<br>
>clustering approach would be more than welcome!<br>
<br>
</div>There's basically one API you implement. If you want to try and implement<br>
that for Terracotta's "next" big thing, you're certainly welcome to.<br>
That's why it's not a migration path from DSO, it's just a totally<br>
different approach.<br>
<br>
My issue with clustering is that if it doesn't solve the maintenance<br>
problem, I'm not very interested in it. By which I mean that if I can't<br>
take down any of my active IdP nodes for maintenance with no service<br>
interruption, I don't understand what it's buying me. I might as well use<br>
active-passive, in which case I can avoid clustering entirely.<br>
<br>
>From a general point of view, memcache is a reasonable substitute for<br>
Terracotta for the time being, in that it adds a process that has to be<br>
running or the system won't work. But that breaks the maintenance property<br>
since I can't patch the memcache host without breaking the IdP.<br>
<br>
The downside of the stateless approach is that you can't implement logout.<br>
Since I believe logout doesn't work anyway, I don't see that as being much<br>
of a tradeoff.<br>
<br>
But V3 needs to provide both alternatives in a maintainable way.<br>
<div class="HOEnZb"><div class="h5"><br>
-- Scott<br>
<br>
<br>
--<br>
To unsubscribe from this list send an email to <a href="mailto:users-unsubscribe@shibboleth.net">users-unsubscribe@shibboleth.net</a><br>
</div></div></blockquote></div><br></div>