[infinispan-issues] [JBoss JIRA] Commented: (ISPN-106) JdbcXXXCacheStore should honor cacheName
Manik Surtani (JIRA)
jira-events at lists.jboss.org
Mon Jul 20 06:25:29 EDT 2009
[ https://jira.jboss.org/jira/browse/ISPN-106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12476826#action_12476826 ]
Manik Surtani commented on ISPN-106:
------------------------------------
I believe different caches should use unique tables, to reduce contention on table-level locks in the DB.
The code snippet you have above should also become a unit test for all cache stores (BaseCacheStoreTest?) to ensure cache stores store stuff in unique locations.
> JdbcXXXCacheStore should honor cacheName
> ----------------------------------------
>
> Key: ISPN-106
> URL: https://jira.jboss.org/jira/browse/ISPN-106
> Project: Infinispan
> Issue Type: Feature Request
> Components: Loaders and Stores
> Affects Versions: 4.0.0.ALPHA5
> Reporter: Klaus Friedel
> Assignee: Manik Surtani
>
> Creating multiple caches from a CacheManager configured with a JdbcXXXCacheStore results in unpredicted behaviour:
> After creating a CacheManger that persists its entries with a JdbcXXXCacheStore one will see strange behaviour like this:
> Cache fooCache = cacheManager.getCache("foo");
> Cache barCache = cacheManager.getCache("bar");
> // single put:
> fooCache.put(42, "Hello World");
> After restart :
> fooCache.contains(42); // true
> barCache.contains(42); // true !!!!
> To circumvent this behaviour the only solution is to create one CacheStore per Cache. So one migth end up using a lot of different CacheManagers.
> The best solution IMHO would be to add sopport for a "cacheName" column to "TableManipulation". So different caches could share the same table.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
More information about the infinispan-issues
mailing list