Hi wychoi, What kind of leak do you have and how do you detect it?
Starting with 1.5, continuous queries support failover, so they maintain the update queues on backup nodes to make sure listener is notified even if primary node crashes. These queues are eventually cleaned up after updates are received by client and are no longer needed, so this should not be the issue. -Val On Tue, Jan 5, 2016 at 9:20 PM, wychoi <[email protected]> wrote: > HI > > We're to examine ignite1.5. > We are being developed using the ContinuousQuery. > > I put the cache in two server nodes, the update notification leak occurred > It did not occur in ignite1.4 > In the ContinuousQuery set RemoteFilter did not use > > Piece of code like this: > > CacheEntryUpdatedListener<AffinityKey, CusTomerObject> > cacheEntryUpdatedListener = new CacheEntryUpdatedListener<AffinityKey, > CusTomerObject>() { > @Override > public void onUpdated( Iterable<CacheEntryEvent<? extends > AffinityKey, ? > extends CusTomerObject>> events) > throws CacheEntryListenerException { > ...... > } > }; > > > ContinuousQuery<AffinityKey, CusTomerObject> query = new > ContinuousQuery<>(); > query.setLocal(true); > query.setTimeInterval(1000); > > ScanQuery scanQuery = new ScanQuery<AffinityKey, CusTomerObject>(new > IgniteBiPredicate<AffinityKey, CusTomerObject>() { > @Override > public boolean apply(AffinityKey key, CusTomerObject val) { > return val.getGrobalKey() > 0; > } > }); > scanQuery.setLocal(true); > scanQuery.setPageSize(1); > > query.setInitialQuery(scanQuery); > query.setLocalListener(cacheEntryUpdatedListener); > .... > > QueryCursor<Cache.Entry<AffinityKey, CusTomerObject>> cursor = > cache.query(query); > > for (Cache.Entry<AffinityKey, CusTomerObject> entry : cursor) { > .... > } > > What should I do to solve this problem ? > > > > > -- > View this message in context: > http://apache-ignite-users.70518.x6.nabble.com/Leaks-in-the-ContinuousQuery-Notification-tp2385.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com. >
