http://tech.puredanger.com/2009/05/01/terracotta-cache-map/ Alex Miller's technical blog Mon, 06 Feb 2012 19:39:50 -0800 hourly 1 http://wordpress.org/?v=3.1.4 http://tech.puredanger.com/2009/05/01/terracotta-cache-map/comment-page-1/#comment-171097 Alex Tue, 21 Jul 2009 10:45:06 +0000 http://tech.puredanger.com/2009/05/01/terracotta-cache-map/#comment-171097 tc-cache will likely be renamed and given some attention in the next release. The Hibernate transactional cache we worked on does buffer up changes but is probably less transactional than you think due to the Hibernate cache provider interface. The cache is really acting as a *listener* to the transactions created in Hibernate, not really as a participant, so there are some possibly strange implications from a generic cache point of view. If you have a caching feature need I would encourage you to file it as a jira request at http://jira.terracotta.org so we can consider it for future incorporation. I don't know of an existing issue that captures the desire for rollback. tc-cache will likely be renamed and given some attention in the next release. The Hibernate transactional cache we worked on does buffer up changes but is probably less transactional than you think due to the Hibernate cache provider interface. The cache is really acting as a *listener* to the transactions created in Hibernate, not really as a participant, so there are some possibly strange implications from a generic cache point of view. If you have a caching feature need I would encourage you to file it as a jira request at http://jira.terracotta.org so we can consider it for future incorporation. I don’t know of an existing issue that captures the desire for rollback.

]]> http://tech.puredanger.com/2009/05/01/terracotta-cache-map/comment-page-1/#comment-171084 Chetan Tue, 21 Jul 2009 07:53:44 +0000 http://tech.puredanger.com/2009/05/01/terracotta-cache-map/#comment-171084 Ok, thats reasonable. We have a use case that tc-cache satisfies...namely the ability for a user to rollback a change as part of a larger transaction, but was under the impression that this was not the strategic direction that was being pursued and that tccache was not in use in production environments. So was looking for other solutions to satisfy this scenario. Ok, thats reasonable. We have a use case that tc-cache satisfies…namely the ability for a user to rollback a change as part of a larger transaction, but was under the impression that this was not the strategic direction that was being pursued and that tccache was not in use in production environments. So was looking for other solutions to satisfy this scenario.

]]> http://tech.puredanger.com/2009/05/01/terracotta-cache-map/comment-page-1/#comment-171023 Alex Mon, 20 Jul 2009 14:53:13 +0000 http://tech.puredanger.com/2009/05/01/terracotta-cache-map/#comment-171023 In 3.1 we are not releasing a transactional cache. We did implement one but decided to focus on read-only and read-write for the 3.1 release. Our general experience with it was that a large number of users use Hibernate in an environment where JTA is not required or available and we were unable to find many users that actually wanted a transactional cache. Most clustered cache users seemed to be using it because that was the only clustered cache level available - those users were perfectly happy to use a coherent read-write cache as that addressed all of the requirements they had. In 3.1 we are not releasing a transactional cache. We did implement one but decided to focus on read-only and read-write for the 3.1 release. Our general experience with it was that a large number of users use Hibernate in an environment where JTA is not required or available and we were unable to find many users that actually wanted a transactional cache. Most clustered cache users seemed to be using it because that was the only clustered cache level available – those users were perfectly happy to use a coherent read-write cache as that addressed all of the requirements they had.

]]> http://tech.puredanger.com/2009/05/01/terracotta-cache-map/comment-page-1/#comment-171022 Anonymous Mon, 20 Jul 2009 14:40:57 +0000 http://tech.puredanger.com/2009/05/01/terracotta-cache-map/#comment-171022 Will the hibernate cache in 3.1 actually be transactional? I had a discussion with Taylor and he mentioned that it would be read-write only. Will the hibernate cache in 3.1 actually be transactional? I had a discussion with Taylor and he mentioned that it would be read-write only.

]]> http://tech.puredanger.com/2009/05/01/terracotta-cache-map/comment-page-1/#comment-163036 Orion Letizi Thu, 07 May 2009 04:18:27 +0000 http://tech.puredanger.com/2009/05/01/terracotta-cache-map/#comment-163036 Awesome writeup, Alex. We need to get this in our docs. Awesome writeup, Alex. We need to get this in our docs.

]]>