Hi,
In the ticket you've shared before, the reproducer has a cache object
creation after the reconnect. This should be done in this case too. The
thing is that the client reconnects to the absolutely different cluster, as
it was in memory and it was fully stopped. To avoid this situation you can
start more than one server node. At the same time, I think that on the
ignite - hibernate integration side, we should add handling of the
reconnect for this scenario.
All cache proxy objects should be recreated after reconnect.
I created ticket for this fix:
https://issues.apache.org/jira/browse/IGNITE-13391
Best Regards,
Evgenii
чт, 27 авг. 2020 г. в 08:48, Tathagata Roy :
> If it helps here is my dependency tree
>
>
>
>
>
>
>
> *From:* Tathagata Roy
> *Sent:* Thursday, August 27, 2020 8:35 AM
> *To:* user
> *Subject:* RE: Hibernate 2nd Level query cache with Ignite
>
>
>
> Thanks for responding @Evgenii
>
>
>
> Attaching the logs for both. There is no signification info in the ignite
> server. In application logs all logs after line 6629 is when application
> was able to reconnect with server after the server restart
>
>
>
> *From:* Evgenii Zhuravlev
> *Sent:* Wednesday, August 26, 2020 4:58 PM
> *To:* user
> *Subject:* Re: Hibernate 2nd Level query cache with Ignite
>
>
>
> Hi,
>
>
>
> Can you please share full logs from client and server nodes?
>
>
>
> Thanks,
>
> Evgenii
>
>
>
> ср, 26 авг. 2020 г. в 14:26, Tathagata Roy >:
>
> Hi,
>
>
>
> I am trying to do a POC on hibernate 2nd level cache with Apache Ignite.
> With this configuration I was able to make it work
>
>
>
> *spring.jpa.properties.hibernate.cache.use_second_level_cache*=
> *true **spring.jpa.properties.hibernate.cache.use_query_cache*=
> *true **spring.jpa.properties.hibernate.generate_statistics*=
> *false **spring.jpa.properties.hibernate.cache.region.factory_class*=
> *org.apache.ignite.cache.hibernate.HibernateRegionFactory *
> *spring.jpa.properties.org.apache.ignite.hibernate.default_access_type*=
> *READ_ONLY*
>
>
>
>
>
> <*dependency*>
> <*groupId*>org.gridgain
> <*artifactId*>ignite-hibernate_5.3
> <*version*>8.7.23
> <*exclusions*>
> <*exclusion*>
> <*groupId*>org.hibernate
> <*artifactId*>hibernate-core
>
>
>
>
>
>
> @Bean
> @ConditionalOnMissingBean
> *public *IgniteConfiguration igniteConfiguration(DiscoverySpi discoverySpi,
> CommunicationSpi communicationSpi) {
> IgniteConfiguration igniteConfiguration = *new *IgniteConfiguration();
> igniteConfiguration.setClientMode(*clientMode*);
> igniteConfiguration.setMetricsLogFrequency(0);
>
> igniteConfiguration.setGridLogger(*new *Slf4jLogger());
>
> igniteConfiguration.setDiscoverySpi(discoverySpi);
> igniteConfiguration.setCommunicationSpi(communicationSpi);
> igniteConfiguration.setFailureDetectionTimeout(*failureDetectionTimeout*);
>
> CacheConfiguration cc = *new *CacheConfiguration<>();
> cc.setName(“Entity1”);
> cc.setCacheMode(CacheMode.*REPLICATED*);
>
>
>
> CacheConfiguration cc1 = *new *CacheConfiguration<>();
> cc1.setName(“*default-query-results-region*”);
> cc1.setCacheMode(CacheMode.*REPLICATED*);
>
>
>
> CacheConfiguration cc2 = *new *CacheConfiguration<>();
> cc2.setName(“*default-update-timestamps-region*”);
> cc2.setCacheMode(CacheMode.*REPLICATED*);
>
> igniteConfiguration.setCacheConfiguration(cc);
>
>
>
> *return *igniteConfiguration;
> }
>
>
>
>
>
>
>
> I am testing this with external ignite node, but if the external ig node
> is restarted , I see the error when trying to access Entity1
>
>
>
> "errorMessage": "class
> org.apache.ignite.internal.processors.cache.CacheStoppedException: Failed
> to perform cache operation (cache is stopped): Entity1; nested exception is
> java.lang.IllegalStateException: class
> org.apache.ignite.internal.processors.cache.CacheStoppedException: Failed
> to perform cache operation (cache is stopped): Entity1",
>
>
>
> It looks like the issue is as reported here ,
>
>
>
>
> https://stackoverflow.com/questions/46053089/ignite-cache-reconnection-issue-cache-is-stopped
>
>
> https://issues.apache.org/jira/browse/IGNITE-5789
>
>
>
>
>
>
> Are there any other way without restaring the client application we can
> make it work?
>
>
