<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br><div><div>On 25 May 2010, at 16:06, Mircea Markus wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br><div><div>On 25 May 2010, at 17:44, Manik Surtani wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div><br>On 25 May 2010, at 15:34, Mircea Markus wrote:<br><br><blockquote type="cite"><br></blockquote><blockquote type="cite">On 25 May 2010, at 17:22, Manik Surtani wrote:<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">On 25 May 2010, at 15:09, Mircea Markus wrote:<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">On 25 May 2010, at 16:26, Manik Surtani wrote:<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">On 25 May 2010, at 14:15, Mircea Markus wrote:<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">But, I'd be concerned about an Infinispan thread that's needed <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">for doing a lot of critical work during a view change getting tied up <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">making a ton of notifications.<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">^^ Yeah that's my concern. &nbsp;If during a rehash we need to stop at every entry that is being moved and issue a notification, that could be costly and really slow down the rehashing process.<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Can't we register an *async* notification listener on ViewChanged ? <br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">You'd still have a *lot* of notifications being queued up for the notification executor since you will have 1 event *per entry* that is moved. &nbsp;<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">this would still happen in the same JVM as this information is needed. Just that we offer this as a service, so that users(AS being one of them) won't have to write this code.<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">No, if the AS does it, it will be one notification (view change) + a scan of known session ids (keys). &nbsp;If we do it, it will involve 1 notification *per key* being migrated.<br></blockquote></blockquote><blockquote type="cite">I don't intend to register more than one ViewChange listener per service. Then AS (or other clients) can register as many keys with that service instance (and consequently with that single listener) and, in that listener to iterate over the keys etc.<br></blockquote><br>You may just register 1 listener, but the event gets dispatched by the rehasher thread, which does:<br></div></blockquote>I was thinking about something totally different.</div><div>1. The NotificationService (NS) is started. Internally it registers a single listener: @ViewChange(async=true)</div><div>2.&nbsp;The user has registers to NS the set of keys for which he wants to be notified on topology changes. Whenever a key is registered, the NS calculates and holds current key address based on the CH.&nbsp;</div><div>3. View change happens, NS is notified with a thread from the async pool</div><div>4. It iterates over the set of registered keys and checks weather the prev address has changed.</div></div></blockquote><div><br></div><div>That's a whole alternate notification system to what we have. &nbsp;:) &nbsp;-1 to supporting 2 such systems.</div><br><blockquote type="cite"><div style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><div><blockquote type="cite"><div><font class="Apple-style-span"><br></font>1. &nbsp;Loop thru all entries. &nbsp;For each entry in the cache:<br>2. &nbsp;&nbsp;&nbsp;&nbsp;Determine if key needs to be moved, given the old topology and the new one after considering a new joiner. &nbsp;If the key needs to be moved:<br>3. &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Add key to digest of entries to be moved.<br>4. &nbsp;Transfer state<br><br>Now if we were to notify listeners of each change, this would happen in the loop, after the condition, around the same time as step 3. &nbsp;Which means a large number of events generated and dispatched (even with just 1 listener). &nbsp;<br><br>Unless you are suggesting that the listener gets a set of keys (extracted from the digest, after the loop) in a single notification event? &nbsp;<br><br><br><blockquote type="cite">As per my understanding, Brian's concern was that the list of keys can be quite large (all the web sessions) and that would delay the rehashing thread - this is now mitigated by using @ViewChange(async=true).<br></blockquote><blockquote type="cite"><blockquote type="cite">The cost of each notification is the construction and initialization of an event object - which includes cloning the invocation context - and placing it on an executor queue.<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Could be thousands, tens of thousands in cases. &nbsp;<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">All other notifications will be severely delayed (depending on your async notifier executor threadpool size)<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">If you have an big enough thread pool (actually at least 2 threads) this should be no issue - and this should be made clear to the user.<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Executor queue size? &nbsp;This could start blocking?<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">--<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Manik Surtani<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><a href="mailto:manik@jboss.org">manik@jboss.org</a><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Lead, Infinispan<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Lead, JBoss Cache<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><a href="http://www.infinispan.org/">http://www.infinispan.org</a><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><a href="http://www.jbosscache.org/">http://www.jbosscache.org</a><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">_______________________________________________<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">infinispan-dev mailing list<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev">https://lists.jboss.org/mailman/listinfo/infinispan-dev</a><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">_______________________________________________<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">infinispan-dev mailing list<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev">https://lists.jboss.org/mailman/listinfo/infinispan-dev</a><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">--<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Manik Surtani<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><a href="mailto:manik@jboss.org">manik@jboss.org</a><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Lead, Infinispan<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">Lead, JBoss Cache<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><a href="http://www.infinispan.org/">http://www.infinispan.org</a><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><a href="http://www.jbosscache.org/">http://www.jbosscache.org</a><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">_______________________________________________<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">infinispan-dev mailing list<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev">https://lists.jboss.org/mailman/listinfo/infinispan-dev</a><br></blockquote></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">_______________________________________________<br></blockquote><blockquote type="cite">infinispan-dev mailing list<br></blockquote><blockquote type="cite"><a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br></blockquote><blockquote type="cite"><a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev">https://lists.jboss.org/mailman/listinfo/infinispan-dev</a><br></blockquote><br>--<br>Manik Surtani<br><a href="mailto:manik@jboss.org">manik@jboss.org</a><br>Lead, Infinispan<br>Lead, JBoss Cache<br><a href="http://www.infinispan.org">http://www.infinispan.org</a><br>http://www.jbosscache.org<br><br><br><br><br><br>_______________________________________________<br>infinispan-dev mailing list<br>infinispan-dev@lists.jboss.org<br>https://lists.jboss.org/mailman/listinfo/infinispan-dev<br></div></blockquote></div><br></div>_______________________________________________<br>infinispan-dev mailing list<br><a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br>https://lists.jboss.org/mailman/listinfo/infinispan-dev</blockquote></div><br><div>
<span class="Apple-style-span" style="border-collapse: separate; color: rgb(0, 0, 0); font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-align: auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-border-horizontal-spacing: 0px; -webkit-border-vertical-spacing: 0px; -webkit-text-decorations-in-effect: none; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0; "><span class="Apple-style-span" style="border-collapse: separate; color: rgb(0, 0, 0); font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-border-horizontal-spacing: 0px; -webkit-border-vertical-spacing: 0px; -webkit-text-decorations-in-effect: none; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; "><div style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><div>--</div><div>Manik Surtani</div><div><a href="mailto:manik@jboss.org">manik@jboss.org</a></div><div>Lead, Infinispan</div><div>Lead, JBoss Cache</div><div><a href="http://www.infinispan.org">http://www.infinispan.org</a></div><div><a href="http://www.jbosscache.org">http://www.jbosscache.org</a></div><div><br></div></div></span><br class="Apple-interchange-newline"></span><br class="Apple-interchange-newline">
</div>
<br></body></html>