store examples
Locked
 
|
https://issues.apache.org/jira/browse/IGNITE-2158
Guys, It seems that CacheDummyStoreExample, CacheJdbcStoreExample and CacheSpringStoreExample do not work correctly in distributed environment after code modifications to demonstrate clear() and withSkipStore() This is an old issue and it is 100% examples' issue, so I suggest removing them from 1.5 and fix and return back to distro in 1.6 Thoughts? --Yakov |
|
If it is an example issue, why can’t we fix the examples? I am not
comfortable removing CacheJdbcStoreExample from the release. On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov <[hidden email]> wrote: > https://issues.apache.org/jira/browse/IGNITE-2158 > > Guys, > > It seems that CacheDummyStoreExample, CacheJdbcStoreExample and > CacheSpringStoreExample do not work correctly in distributed environment > after code modifications to demonstrate clear() and withSkipStore() > > This is an old issue and it is 100% examples' issue, so I suggest removing > them from 1.5 and fix and return back to distro in 1.6 > > Thoughts? > > --Yakov > |
|
|
We still have very advanced and cool CacheBinaryAutoStoreExample and it
works properly --Yakov 2015-12-15 18:11 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > If it is an example issue, why can’t we fix the examples? I am not > comfortable removing CacheJdbcStoreExample from the release. > > On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov <[hidden email]> > wrote: > > > https://issues.apache.org/jira/browse/IGNITE-2158 > > > > Guys, > > > > It seems that CacheDummyStoreExample, CacheJdbcStoreExample and > > CacheSpringStoreExample do not work correctly in distributed environment > > after code modifications to demonstrate clear() and withSkipStore() > > > > This is an old issue and it is 100% examples' issue, so I suggest > removing > > them from 1.5 and fix and return back to distro in 1.6 > > > > Thoughts? > > > > --Yakov > > > |
|
|
I think we should fix all the examples. What has changed that makes these
examples so difficult to fix? I really do not like removing the examples that have been continuously shipped in every release. If it is so difficult for us to upgrade them to 1.5, how difficult will it be for our users? D. On Tue, Dec 15, 2015 at 7:24 AM, Yakov Zhdanov <[hidden email]> wrote: > We still have very advanced and cool CacheBinaryAutoStoreExample and it > works properly > > --Yakov > > 2015-12-15 18:11 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > If it is an example issue, why can’t we fix the examples? I am not > > comfortable removing CacheJdbcStoreExample from the release. > > > > On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov <[hidden email]> > > wrote: > > > > > https://issues.apache.org/jira/browse/IGNITE-2158 > > > > > > Guys, > > > > > > It seems that CacheDummyStoreExample, CacheJdbcStoreExample and > > > CacheSpringStoreExample do not work correctly in distributed > environment > > > after code modifications to demonstrate clear() and withSkipStore() > > > > > > This is an old issue and it is 100% examples' issue, so I suggest > > removing > > > them from 1.5 and fix and return back to distro in 1.6 > > > > > > Thoughts? > > > > > > --Yakov > > > > > > |
|
|
As far as I understand, the issue is that CacheJdbcStoreExample
and CacheSpringStoreExample use in-memory H2 database, so the cache store is not shared among nodes. What if we just switch to a file-based DB? It looks like a simple change to me. And I would remove CacheDummyStoreExample. It looks weird and doesn't demonstrate anything that is not shown in other store examples. -Val On Tue, Dec 15, 2015 at 2:03 PM, Dmitriy Setrakyan <[hidden email]> wrote: > I think we should fix all the examples. What has changed that makes these > examples so difficult to fix? > > I really do not like removing the examples that have been continuously > shipped in every release. If it is so difficult for us to upgrade them to > 1.5, how difficult will it be for our users? > > D. > > On Tue, Dec 15, 2015 at 7:24 AM, Yakov Zhdanov <[hidden email]> > wrote: > > > We still have very advanced and cool CacheBinaryAutoStoreExample and it > > works properly > > > > --Yakov > > > > 2015-12-15 18:11 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > > > If it is an example issue, why can’t we fix the examples? I am not > > > comfortable removing CacheJdbcStoreExample from the release. > > > > > > On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov <[hidden email]> > > > wrote: > > > > > > > https://issues.apache.org/jira/browse/IGNITE-2158 > > > > > > > > Guys, > > > > > > > > It seems that CacheDummyStoreExample, CacheJdbcStoreExample and > > > > CacheSpringStoreExample do not work correctly in distributed > > environment > > > > after code modifications to demonstrate clear() and withSkipStore() > > > > > > > > This is an old issue and it is 100% examples' issue, so I suggest > > > removing > > > > them from 1.5 and fix and return back to distro in 1.6 > > > > > > > > Thoughts? > > > > > > > > --Yakov > > > > > > > > > > |
|
|
If the problem is the H2 in-memory store not being shared between multiple
Ignite nodes, then why not switch to using H2 disk store? On Tue, Dec 15, 2015 at 4:11 PM, Valentin Kulichenko < [hidden email]> wrote: > As far as I understand, the issue is that CacheJdbcStoreExample > and CacheSpringStoreExample use in-memory H2 database, so the cache store > is not shared among nodes. What if we just switch to a file-based DB? It > looks like a simple change to me. > > And I would remove CacheDummyStoreExample. It looks weird and doesn't > demonstrate anything that is not shown in other store examples. > > -Val > > On Tue, Dec 15, 2015 at 2:03 PM, Dmitriy Setrakyan <[hidden email]> > wrote: > > > I think we should fix all the examples. What has changed that makes these > > examples so difficult to fix? > > > > I really do not like removing the examples that have been continuously > > shipped in every release. If it is so difficult for us to upgrade them to > > 1.5, how difficult will it be for our users? > > > > D. > > > > On Tue, Dec 15, 2015 at 7:24 AM, Yakov Zhdanov <[hidden email]> > > wrote: > > > > > We still have very advanced and cool CacheBinaryAutoStoreExample and it > > > works properly > > > > > > --Yakov > > > > > > 2015-12-15 18:11 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > > > > > If it is an example issue, why can’t we fix the examples? I am not > > > > comfortable removing CacheJdbcStoreExample from the release. > > > > > > > > On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov <[hidden email]> > > > > wrote: > > > > > > > > > https://issues.apache.org/jira/browse/IGNITE-2158 > > > > > > > > > > Guys, > > > > > > > > > > It seems that CacheDummyStoreExample, CacheJdbcStoreExample and > > > > > CacheSpringStoreExample do not work correctly in distributed > > > environment > > > > > after code modifications to demonstrate clear() and withSkipStore() > > > > > > > > > > This is an old issue and it is 100% examples' issue, so I suggest > > > > removing > > > > > them from 1.5 and fix and return back to distro in 1.6 > > > > > > > > > > Thoughts? > > > > > > > > > > --Yakov > > > > > > > > > > > > > > > |
|
|
I think mentioned examples could be easily fixed by switching from use of
H2 started from example to o.a.i.examples.util.DbH2ServerStartup. This will start full fledged H2 server with in-memory db and all nodes will work with it. >> Ignite nodes, then why not switch to using H2 disk store? Dima, this would not work, because file will be locked by first started process and other will failed. I think we should go as I mentioned above. On Wed, Dec 16, 2015 at 7:23 AM, Dmitriy Setrakyan <[hidden email]> wrote: > If the problem is the H2 in-memory store not being shared between multiple > Ignite nodes, then why not switch to using H2 disk store? > > On Tue, Dec 15, 2015 at 4:11 PM, Valentin Kulichenko < > [hidden email]> wrote: > > > As far as I understand, the issue is that CacheJdbcStoreExample > > and CacheSpringStoreExample use in-memory H2 database, so the cache store > > is not shared among nodes. What if we just switch to a file-based DB? It > > looks like a simple change to me. > > > > And I would remove CacheDummyStoreExample. It looks weird and doesn't > > demonstrate anything that is not shown in other store examples. > > > > -Val > > > > On Tue, Dec 15, 2015 at 2:03 PM, Dmitriy Setrakyan < > [hidden email]> > > wrote: > > > > > I think we should fix all the examples. What has changed that makes > these > > > examples so difficult to fix? > > > > > > I really do not like removing the examples that have been continuously > > > shipped in every release. If it is so difficult for us to upgrade them > to > > > 1.5, how difficult will it be for our users? > > > > > > D. > > > > > > On Tue, Dec 15, 2015 at 7:24 AM, Yakov Zhdanov <[hidden email]> > > > wrote: > > > > > > > We still have very advanced and cool CacheBinaryAutoStoreExample and > it > > > > works properly > > > > > > > > --Yakov > > > > > > > > 2015-12-15 18:11 GMT+03:00 Dmitriy Setrakyan <[hidden email] > >: > > > > > > > > > If it is an example issue, why can’t we fix the examples? I am not > > > > > comfortable removing CacheJdbcStoreExample from the release. > > > > > > > > > > On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov < > [hidden email]> > > > > > wrote: > > > > > > > > > > > https://issues.apache.org/jira/browse/IGNITE-2158 > > > > > > > > > > > > Guys, > > > > > > > > > > > > It seems that CacheDummyStoreExample, CacheJdbcStoreExample and > > > > > > CacheSpringStoreExample do not work correctly in distributed > > > > environment > > > > > > after code modifications to demonstrate clear() and > withSkipStore() > > > > > > > > > > > > This is an old issue and it is 100% examples' issue, so I suggest > > > > > removing > > > > > > them from 1.5 and fix and return back to distro in 1.6 > > > > > > > > > > > > Thoughts? > > > > > > > > > > > > --Yakov > > > > > > > > > > > > > > > > > > > > > -- Alexey Kuznetsov GridGain Systems www.gridgain.com |
|
|
Got it. Yakov, I think this will work without any coding change. Can you
confirm? On Tue, Dec 15, 2015 at 6:30 PM, Alexey Kuznetsov <[hidden email]> wrote: > I think mentioned examples could be easily fixed by switching from use of > H2 started from example to o.a.i.examples.util.DbH2ServerStartup. > This will start full fledged H2 server with in-memory db and all nodes will > work with it. > > >> Ignite nodes, then why not switch to using H2 disk store? > Dima, this would not work, because file will be locked by first started > process and other will failed. > > I think we should go as I mentioned above. > > > On Wed, Dec 16, 2015 at 7:23 AM, Dmitriy Setrakyan <[hidden email]> > wrote: > > > If the problem is the H2 in-memory store not being shared between > multiple > > Ignite nodes, then why not switch to using H2 disk store? > > > > On Tue, Dec 15, 2015 at 4:11 PM, Valentin Kulichenko < > > [hidden email]> wrote: > > > > > As far as I understand, the issue is that CacheJdbcStoreExample > > > and CacheSpringStoreExample use in-memory H2 database, so the cache > store > > > is not shared among nodes. What if we just switch to a file-based DB? > It > > > looks like a simple change to me. > > > > > > And I would remove CacheDummyStoreExample. It looks weird and doesn't > > > demonstrate anything that is not shown in other store examples. > > > > > > -Val > > > > > > On Tue, Dec 15, 2015 at 2:03 PM, Dmitriy Setrakyan < > > [hidden email]> > > > wrote: > > > > > > > I think we should fix all the examples. What has changed that makes > > these > > > > examples so difficult to fix? > > > > > > > > I really do not like removing the examples that have been > continuously > > > > shipped in every release. If it is so difficult for us to upgrade > them > > to > > > > 1.5, how difficult will it be for our users? > > > > > > > > D. > > > > > > > > On Tue, Dec 15, 2015 at 7:24 AM, Yakov Zhdanov <[hidden email]> > > > > wrote: > > > > > > > > > We still have very advanced and cool CacheBinaryAutoStoreExample > and > > it > > > > > works properly > > > > > > > > > > --Yakov > > > > > > > > > > 2015-12-15 18:11 GMT+03:00 Dmitriy Setrakyan < > [hidden email] > > >: > > > > > > > > > > > If it is an example issue, why can’t we fix the examples? I am > not > > > > > > comfortable removing CacheJdbcStoreExample from the release. > > > > > > > > > > > > On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov < > > [hidden email]> > > > > > > wrote: > > > > > > > > > > > > > https://issues.apache.org/jira/browse/IGNITE-2158 > > > > > > > > > > > > > > Guys, > > > > > > > > > > > > > > It seems that CacheDummyStoreExample, CacheJdbcStoreExample and > > > > > > > CacheSpringStoreExample do not work correctly in distributed > > > > > environment > > > > > > > after code modifications to demonstrate clear() and > > withSkipStore() > > > > > > > > > > > > > > This is an old issue and it is 100% examples' issue, so I > suggest > > > > > > removing > > > > > > > them from 1.5 and fix and return back to distro in 1.6 > > > > > > > > > > > > > > Thoughts? > > > > > > > > > > > > > > --Yakov > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > Alexey Kuznetsov > GridGain Systems > www.gridgain.com > |
|
|
Guys,
I refactored JDBC store example to use h2 server startup. I decided to remove spring store example since code will be too complex after fixing all issues with transactions along with jdbcTemplate (current code absolutely ignores txs) + I removed dummy example as per Val's suggestion. --Yakov 2015-12-16 9:23 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > Got it. Yakov, I think this will work without any coding change. Can you > confirm? > > On Tue, Dec 15, 2015 at 6:30 PM, Alexey Kuznetsov <[hidden email] > > > wrote: > > > I think mentioned examples could be easily fixed by switching from use of > > H2 started from example to o.a.i.examples.util.DbH2ServerStartup. > > This will start full fledged H2 server with in-memory db and all nodes > will > > work with it. > > > > >> Ignite nodes, then why not switch to using H2 disk store? > > Dima, this would not work, because file will be locked by first started > > process and other will failed. > > > > I think we should go as I mentioned above. > > > > > > On Wed, Dec 16, 2015 at 7:23 AM, Dmitriy Setrakyan < > [hidden email]> > > wrote: > > > > > If the problem is the H2 in-memory store not being shared between > > multiple > > > Ignite nodes, then why not switch to using H2 disk store? > > > > > > On Tue, Dec 15, 2015 at 4:11 PM, Valentin Kulichenko < > > > [hidden email]> wrote: > > > > > > > As far as I understand, the issue is that CacheJdbcStoreExample > > > > and CacheSpringStoreExample use in-memory H2 database, so the cache > > store > > > > is not shared among nodes. What if we just switch to a file-based DB? > > It > > > > looks like a simple change to me. > > > > > > > > And I would remove CacheDummyStoreExample. It looks weird and doesn't > > > > demonstrate anything that is not shown in other store examples. > > > > > > > > -Val > > > > > > > > On Tue, Dec 15, 2015 at 2:03 PM, Dmitriy Setrakyan < > > > [hidden email]> > > > > wrote: > > > > > > > > > I think we should fix all the examples. What has changed that makes > > > these > > > > > examples so difficult to fix? > > > > > > > > > > I really do not like removing the examples that have been > > continuously > > > > > shipped in every release. If it is so difficult for us to upgrade > > them > > > to > > > > > 1.5, how difficult will it be for our users? > > > > > > > > > > D. > > > > > > > > > > On Tue, Dec 15, 2015 at 7:24 AM, Yakov Zhdanov < > [hidden email]> > > > > > wrote: > > > > > > > > > > > We still have very advanced and cool CacheBinaryAutoStoreExample > > and > > > it > > > > > > works properly > > > > > > > > > > > > --Yakov > > > > > > > > > > > > 2015-12-15 18:11 GMT+03:00 Dmitriy Setrakyan < > > [hidden email] > > > >: > > > > > > > > > > > > > If it is an example issue, why can’t we fix the examples? I am > > not > > > > > > > comfortable removing CacheJdbcStoreExample from the release. > > > > > > > > > > > > > > On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov < > > > [hidden email]> > > > > > > > wrote: > > > > > > > > > > > > > > > https://issues.apache.org/jira/browse/IGNITE-2158 > > > > > > > > > > > > > > > > Guys, > > > > > > > > > > > > > > > > It seems that CacheDummyStoreExample, CacheJdbcStoreExample > and > > > > > > > > CacheSpringStoreExample do not work correctly in distributed > > > > > > environment > > > > > > > > after code modifications to demonstrate clear() and > > > withSkipStore() > > > > > > > > > > > > > > > > This is an old issue and it is 100% examples' issue, so I > > suggest > > > > > > > removing > > > > > > > > them from 1.5 and fix and return back to distro in 1.6 > > > > > > > > > > > > > > > > Thoughts? > > > > > > > > > > > > > > > > --Yakov > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > Alexey Kuznetsov > > GridGain Systems > > www.gridgain.com > > > |
|
|
Yakov,
Code in Spring example was not ignoring txs. There was CacheJdbcStoreSessionListener which handles this automatically. That's how Spring works - it enlists all calls to jdbcTemplate into transaction if it was opened by tx manager in this thread. -Val On Wed, Dec 16, 2015 at 6:35 AM, Yakov Zhdanov <[hidden email]> wrote: > Guys, > > I refactored JDBC store example to use h2 server startup. > > I decided to remove spring store example since code will be too complex > after fixing all issues with transactions along with jdbcTemplate (current > code absolutely ignores txs) > + I removed dummy example as per Val's suggestion. > > --Yakov > > 2015-12-16 9:23 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > Got it. Yakov, I think this will work without any coding change. Can you > > confirm? > > > > On Tue, Dec 15, 2015 at 6:30 PM, Alexey Kuznetsov < > [hidden email] > > > > > wrote: > > > > > I think mentioned examples could be easily fixed by switching from use > of > > > H2 started from example to o.a.i.examples.util.DbH2ServerStartup. > > > This will start full fledged H2 server with in-memory db and all nodes > > will > > > work with it. > > > > > > >> Ignite nodes, then why not switch to using H2 disk store? > > > Dima, this would not work, because file will be locked by first started > > > process and other will failed. > > > > > > I think we should go as I mentioned above. > > > > > > > > > On Wed, Dec 16, 2015 at 7:23 AM, Dmitriy Setrakyan < > > [hidden email]> > > > wrote: > > > > > > > If the problem is the H2 in-memory store not being shared between > > > multiple > > > > Ignite nodes, then why not switch to using H2 disk store? > > > > > > > > On Tue, Dec 15, 2015 at 4:11 PM, Valentin Kulichenko < > > > > [hidden email]> wrote: > > > > > > > > > As far as I understand, the issue is that CacheJdbcStoreExample > > > > > and CacheSpringStoreExample use in-memory H2 database, so the cache > > > store > > > > > is not shared among nodes. What if we just switch to a file-based > DB? > > > It > > > > > looks like a simple change to me. > > > > > > > > > > And I would remove CacheDummyStoreExample. It looks weird and > doesn't > > > > > demonstrate anything that is not shown in other store examples. > > > > > > > > > > -Val > > > > > > > > > > On Tue, Dec 15, 2015 at 2:03 PM, Dmitriy Setrakyan < > > > > [hidden email]> > > > > > wrote: > > > > > > > > > > > I think we should fix all the examples. What has changed that > makes > > > > these > > > > > > examples so difficult to fix? > > > > > > > > > > > > I really do not like removing the examples that have been > > > continuously > > > > > > shipped in every release. If it is so difficult for us to upgrade > > > them > > > > to > > > > > > 1.5, how difficult will it be for our users? > > > > > > > > > > > > D. > > > > > > > > > > > > On Tue, Dec 15, 2015 at 7:24 AM, Yakov Zhdanov < > > [hidden email]> > > > > > > wrote: > > > > > > > > > > > > > We still have very advanced and cool > CacheBinaryAutoStoreExample > > > and > > > > it > > > > > > > works properly > > > > > > > > > > > > > > --Yakov > > > > > > > > > > > > > > 2015-12-15 18:11 GMT+03:00 Dmitriy Setrakyan < > > > [hidden email] > > > > >: > > > > > > > > > > > > > > > If it is an example issue, why can’t we fix the examples? I > am > > > not > > > > > > > > comfortable removing CacheJdbcStoreExample from the release. > > > > > > > > > > > > > > > > On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov < > > > > [hidden email]> > > > > > > > > wrote: > > > > > > > > > > > > > > > > > https://issues.apache.org/jira/browse/IGNITE-2158 > > > > > > > > > > > > > > > > > > Guys, > > > > > > > > > > > > > > > > > > It seems that CacheDummyStoreExample, CacheJdbcStoreExample > > and > > > > > > > > > CacheSpringStoreExample do not work correctly in > distributed > > > > > > > environment > > > > > > > > > after code modifications to demonstrate clear() and > > > > withSkipStore() > > > > > > > > > > > > > > > > > > This is an old issue and it is 100% examples' issue, so I > > > suggest > > > > > > > > removing > > > > > > > > > them from 1.5 and fix and return back to distro in 1.6 > > > > > > > > > > > > > > > > > > Thoughts? > > > > > > > > > > > > > > > > > > --Yakov > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > Alexey Kuznetsov > > > GridGain Systems > > > www.gridgain.com > > > > > > |
|
|
Can we please add this example back?
On Wed, Dec 16, 2015 at 10:45 PM, Valentin Kulichenko < [hidden email]> wrote: > Yakov, > > Code in Spring example was not ignoring txs. There > was CacheJdbcStoreSessionListener which handles this automatically. That's > how Spring works - it enlists all calls to jdbcTemplate into transaction if > it was opened by tx manager in this thread. > > -Val > > On Wed, Dec 16, 2015 at 6:35 AM, Yakov Zhdanov <[hidden email]> > wrote: > > > Guys, > > > > I refactored JDBC store example to use h2 server startup. > > > > I decided to remove spring store example since code will be too complex > > after fixing all issues with transactions along with jdbcTemplate > (current > > code absolutely ignores txs) > > + I removed dummy example as per Val's suggestion. > > > > --Yakov > > > > 2015-12-16 9:23 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > > > Got it. Yakov, I think this will work without any coding change. Can > you > > > confirm? > > > > > > On Tue, Dec 15, 2015 at 6:30 PM, Alexey Kuznetsov < > > [hidden email] > > > > > > > wrote: > > > > > > > I think mentioned examples could be easily fixed by switching from > use > > of > > > > H2 started from example to o.a.i.examples.util.DbH2ServerStartup. > > > > This will start full fledged H2 server with in-memory db and all > nodes > > > will > > > > work with it. > > > > > > > > >> Ignite nodes, then why not switch to using H2 disk store? > > > > Dima, this would not work, because file will be locked by first > started > > > > process and other will failed. > > > > > > > > I think we should go as I mentioned above. > > > > > > > > > > > > On Wed, Dec 16, 2015 at 7:23 AM, Dmitriy Setrakyan < > > > [hidden email]> > > > > wrote: > > > > > > > > > If the problem is the H2 in-memory store not being shared between > > > > multiple > > > > > Ignite nodes, then why not switch to using H2 disk store? > > > > > > > > > > On Tue, Dec 15, 2015 at 4:11 PM, Valentin Kulichenko < > > > > > [hidden email]> wrote: > > > > > > > > > > > As far as I understand, the issue is that CacheJdbcStoreExample > > > > > > and CacheSpringStoreExample use in-memory H2 database, so the > cache > > > > store > > > > > > is not shared among nodes. What if we just switch to a file-based > > DB? > > > > It > > > > > > looks like a simple change to me. > > > > > > > > > > > > And I would remove CacheDummyStoreExample. It looks weird and > > doesn't > > > > > > demonstrate anything that is not shown in other store examples. > > > > > > > > > > > > -Val > > > > > > > > > > > > On Tue, Dec 15, 2015 at 2:03 PM, Dmitriy Setrakyan < > > > > > [hidden email]> > > > > > > wrote: > > > > > > > > > > > > > I think we should fix all the examples. What has changed that > > makes > > > > > these > > > > > > > examples so difficult to fix? > > > > > > > > > > > > > > I really do not like removing the examples that have been > > > > continuously > > > > > > > shipped in every release. If it is so difficult for us to > upgrade > > > > them > > > > > to > > > > > > > 1.5, how difficult will it be for our users? > > > > > > > > > > > > > > D. > > > > > > > > > > > > > > On Tue, Dec 15, 2015 at 7:24 AM, Yakov Zhdanov < > > > [hidden email]> > > > > > > > wrote: > > > > > > > > > > > > > > > We still have very advanced and cool > > CacheBinaryAutoStoreExample > > > > and > > > > > it > > > > > > > > works properly > > > > > > > > > > > > > > > > --Yakov > > > > > > > > > > > > > > > > 2015-12-15 18:11 GMT+03:00 Dmitriy Setrakyan < > > > > [hidden email] > > > > > >: > > > > > > > > > > > > > > > > > If it is an example issue, why can’t we fix the examples? I > > am > > > > not > > > > > > > > > comfortable removing CacheJdbcStoreExample from the > release. > > > > > > > > > > > > > > > > > > On Tue, Dec 15, 2015 at 4:02 AM, Yakov Zhdanov < > > > > > [hidden email]> > > > > > > > > > wrote: > > > > > > > > > > > > > > > > > > > https://issues.apache.org/jira/browse/IGNITE-2158 > > > > > > > > > > > > > > > > > > > > Guys, > > > > > > > > > > > > > > > > > > > > It seems that CacheDummyStoreExample, > CacheJdbcStoreExample > > > and > > > > > > > > > > CacheSpringStoreExample do not work correctly in > > distributed > > > > > > > > environment > > > > > > > > > > after code modifications to demonstrate clear() and > > > > > withSkipStore() > > > > > > > > > > > > > > > > > > > > This is an old issue and it is 100% examples' issue, so I > > > > suggest > > > > > > > > > removing > > > > > > > > > > them from 1.5 and fix and return back to distro in 1.6 > > > > > > > > > > > > > > > > > > > > Thoughts? > > > > > > > > > > > > > > > > > > > > --Yakov > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > Alexey Kuznetsov > > > > GridGain Systems > > > > www.gridgain.com > > > > > > > > > > |
|
|
No, I don't want to add it back since example does not work.
Val, I have just very thoroughly rechecked this and it seems you are far from being right. I see that connection used in ignite store listener and connection (and, moreover, sometimes 2 connections - in case of update-insert sequence) used by jdbc template are different. And each connection gets requested and closed by jdbc template on per operation basis. All you talk about can be achieved with explicitly configured tx manager. I don't want to do it since it is 100% overkill for this example. So, I prefer not to add it back - we have 2 store examples more. --Yakov 2015-12-17 11:22 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > Can we please add this example back? > > On Wed, Dec 16, 2015 at 10:45 PM, Valentin Kulichenko < > [hidden email]> wrote: > > > Yakov, > > > > Code in Spring example was not ignoring txs. There > > was CacheJdbcStoreSessionListener which handles this automatically. > That's > > how Spring works - it enlists all calls to jdbcTemplate into transaction > if > > it was opened by tx manager in this thread. > > > > -Val > |
|
|
Yakov,
And what would be the right way to fix this example? Remember, this example exists because we have gotten questions about integrating with spring transactions from our users. D. On Thu, Dec 17, 2015 at 1:00 AM, Yakov Zhdanov <[hidden email]> wrote: > No, I don't want to add it back since example does not work. > > Val, I have just very thoroughly rechecked this and it seems you are far > from being right. > > I see that connection used in ignite store listener and connection (and, > moreover, sometimes 2 connections - in case of update-insert sequence) used > by jdbc template are different. And each connection gets requested and > closed by jdbc template on per operation basis. > > All you talk about can be achieved with explicitly configured tx manager. I > don't want to do it since it is 100% overkill for this example. > > So, I prefer not to add it back - we have 2 store examples more. > > --Yakov > > 2015-12-17 11:22 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > Can we please add this example back? > > > > On Wed, Dec 16, 2015 at 10:45 PM, Valentin Kulichenko < > > [hidden email]> wrote: > > > > > Yakov, > > > > > > Code in Spring example was not ignoring txs. There > > > was CacheJdbcStoreSessionListener which handles this automatically. > > That's > > > how Spring works - it enlists all calls to jdbcTemplate into > transaction > > if > > > it was opened by tx manager in this thread. > > > > > > -Val > > > |
|
|
Example, in state it has been removed in, does not add any value, but only
confusion. I think properly configured Spring stuff will make the trick. Most likely, we will also need spring-enabled store listener. Val, can you please investigate and file a ticket with description? --Yakov 2015-12-17 12:10 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > Yakov, > > And what would be the right way to fix this example? Remember, this example > exists because we have gotten questions about integrating with spring > transactions from our users. > > D. > > On Thu, Dec 17, 2015 at 1:00 AM, Yakov Zhdanov <[hidden email]> > wrote: > > > No, I don't want to add it back since example does not work. > > > > Val, I have just very thoroughly rechecked this and it seems you are far > > from being right. > > > > I see that connection used in ignite store listener and connection (and, > > moreover, sometimes 2 connections - in case of update-insert sequence) > used > > by jdbc template are different. And each connection gets requested and > > closed by jdbc template on per operation basis. > > > > All you talk about can be achieved with explicitly configured tx > manager. I > > don't want to do it since it is 100% overkill for this example. > > > > So, I prefer not to add it back - we have 2 store examples more. > > > > --Yakov > > > > 2015-12-17 11:22 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > > > Can we please add this example back? > > > > > > On Wed, Dec 16, 2015 at 10:45 PM, Valentin Kulichenko < > > > [hidden email]> wrote: > > > > > > > Yakov, > > > > > > > > Code in Spring example was not ignoring txs. There > > > > was CacheJdbcStoreSessionListener which handles this automatically. > > > That's > > > > how Spring works - it enlists all calls to jdbcTemplate into > > transaction > > > if > > > > it was opened by tx manager in this thread. > > > > > > > > -Val > > > > > > |
|
|
Yakov,
We already have CacheSpringStoreSessionListener, but it was not used in the example (accidentally, I guess). I restored the example, made the fix and switched it to DbH2ServerStartup. Now it seems to work correctly. I also made similar changes to CacheHibernateStoreExample. Can you please review my changes? -Val On Thu, Dec 17, 2015 at 1:39 AM, Yakov Zhdanov <[hidden email]> wrote: > Example, in state it has been removed in, does not add any value, but only > confusion. > > I think properly configured Spring stuff will make the trick. Most likely, > we will also need spring-enabled store listener. > > Val, can you please investigate and file a ticket with description? > > --Yakov > > 2015-12-17 12:10 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > Yakov, > > > > And what would be the right way to fix this example? Remember, this > example > > exists because we have gotten questions about integrating with spring > > transactions from our users. > > > > D. > > > > On Thu, Dec 17, 2015 at 1:00 AM, Yakov Zhdanov <[hidden email]> > > wrote: > > > > > No, I don't want to add it back since example does not work. > > > > > > Val, I have just very thoroughly rechecked this and it seems you are > far > > > from being right. > > > > > > I see that connection used in ignite store listener and connection > (and, > > > moreover, sometimes 2 connections - in case of update-insert sequence) > > used > > > by jdbc template are different. And each connection gets requested and > > > closed by jdbc template on per operation basis. > > > > > > All you talk about can be achieved with explicitly configured tx > > manager. I > > > don't want to do it since it is 100% overkill for this example. > > > > > > So, I prefer not to add it back - we have 2 store examples more. > > > > > > --Yakov > > > > > > 2015-12-17 11:22 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > > > > > Can we please add this example back? > > > > > > > > On Wed, Dec 16, 2015 at 10:45 PM, Valentin Kulichenko < > > > > [hidden email]> wrote: > > > > > > > > > Yakov, > > > > > > > > > > Code in Spring example was not ignoring txs. There > > > > > was CacheJdbcStoreSessionListener which handles this automatically. > > > > That's > > > > > how Spring works - it enlists all calls to jdbcTemplate into > > > transaction > > > > if > > > > > it was opened by tx manager in this thread. > > > > > > > > > > -Val > > > > > > > > > > |
|
|
Thanks Valentin!
On Thu, Dec 17, 2015 at 4:07 PM, Valentin Kulichenko < [hidden email]> wrote: > Yakov, > > We already have CacheSpringStoreSessionListener, but it was not used in the > example (accidentally, I guess). I restored the example, made the fix and > switched it to DbH2ServerStartup. Now it seems to work correctly. > > I also made similar changes to CacheHibernateStoreExample. > > Can you please review my changes? > > -Val > > On Thu, Dec 17, 2015 at 1:39 AM, Yakov Zhdanov <[hidden email]> > wrote: > > > Example, in state it has been removed in, does not add any value, but > only > > confusion. > > > > I think properly configured Spring stuff will make the trick. Most > likely, > > we will also need spring-enabled store listener. > > > > Val, can you please investigate and file a ticket with description? > > > > --Yakov > > > > 2015-12-17 12:10 GMT+03:00 Dmitriy Setrakyan <[hidden email]>: > > > > > Yakov, > > > > > > And what would be the right way to fix this example? Remember, this > > example > > > exists because we have gotten questions about integrating with spring > > > transactions from our users. > > > > > > D. > > > > > > On Thu, Dec 17, 2015 at 1:00 AM, Yakov Zhdanov <[hidden email]> > > > wrote: > > > > > > > No, I don't want to add it back since example does not work. > > > > > > > > Val, I have just very thoroughly rechecked this and it seems you are > > far > > > > from being right. > > > > > > > > I see that connection used in ignite store listener and connection > > (and, > > > > moreover, sometimes 2 connections - in case of update-insert > sequence) > > > used > > > > by jdbc template are different. And each connection gets requested > and > > > > closed by jdbc template on per operation basis. > > > > > > > > All you talk about can be achieved with explicitly configured tx > > > manager. I > > > > don't want to do it since it is 100% overkill for this example. > > > > > > > > So, I prefer not to add it back - we have 2 store examples more. > > > > > > > > --Yakov > > > > > > > > 2015-12-17 11:22 GMT+03:00 Dmitriy Setrakyan <[hidden email] > >: > > > > > > > > > Can we please add this example back? > > > > > > > > > > On Wed, Dec 16, 2015 at 10:45 PM, Valentin Kulichenko < > > > > > [hidden email]> wrote: > > > > > > > > > > > Yakov, > > > > > > > > > > > > Code in Spring example was not ignoring txs. There > > > > > > was CacheJdbcStoreSessionListener which handles this > automatically. > > > > > That's > > > > > > how Spring works - it enlists all calls to jdbcTemplate into > > > > transaction > > > > > if > > > > > > it was opened by tx manager in this thread. > > > > > > > > > > > > -Val > > > > > > > > > > > > > > > |
|
|
In reply to this post by Valentin Kulichenko
Reviewed, looks good.
On 12/18/2015 3:07 AM, Valentin Kulichenko wrote: > Can you please review my changes? |
«
Return to Apache Ignite Developers - Legacy Mail Archive
|
1 view|%1 views
| Free forum by Nabble | Edit this page |