Author |
Message |
|
Was their a restart of some sort?
|
|
|
Thanks, We'll check it out.
|
|
|
In Terracotta 3.6.1 ehcache 2.5.1 you can turn off the client side cache completely. Maybe give that a try?
|
|
|
Hi,
Package up a sample that demonstrates the issue and we'll take a look. I haven't heard anything like this before so I'm curious.
|
|
|
Right, it will get retrieved but only the data that is needed.
|
|
|
L2 should give you milli-second latency vs micro-second latency for things cached in the L1.
|
|
|
Yeah, entries on nodes that are stale are notified so they can be invalidated.
In answer to your other question. Yes, you can turn off the local cache.
|
|
|
Hi, You pretty much nailed it with 1 exception. The L2 doesn't push data to L1's when a person tries to retrieve data from the second L1 it will get faulted from the L2 the first time. That said, if the data is changed the data will be kept fresh in the cluster.
Does that make sense?
Cheers,
Steve
|
|
|
Definitely upgrade to 2.5.1 as we fixed a bunch of hibernate issues in that release.
|
|
|
what version of Ehcache are you using? I know a bunch of hibernate issues were resolved in 3.6.1 with ehcache 2.5.1.
|
|
|
I'm sorry but I'm not clear what the question is here?
|
|
|
Hi, Can you throw a JIRA in on this and we'll take a look.
jira.terracotta.org
cheers,
Steve
|
|
|
Did you try increasing your Perm Gen space in your JVM?
|
|
|
Yes, we are actively moving to Java 7. Our trunk nightly builds are jdk 7 compatible.
|
|
|
Glad you figured it out!
|
|
|