[
https://issues.apache.org/jira/browse/GEODE-5181?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Eric Shu updated GEODE-5181:
----------------------------
Affects Version/s: 1.1.0
1.1.1
1.2.0
1.3.0
1.4.0
1.5.0
1.6.0
> A get operation in a transaction on a persistent region on a restarted server
> could fail
> ----------------------------------------------------------------------------------------
>
> Key: GEODE-5181
> URL: https://issues.apache.org/jira/browse/GEODE-5181
> Project: Geode
> Issue Type: Bug
> Components: persistence, transactions
> Affects Versions: 1.1.0, 1.1.1, 1.2.0, 1.3.0, 1.4.0, 1.5.0, 1.6.0
> Reporter: Eric Shu
> Priority: Major
>
> When trying to fix GEODE-5173, we find out that transaction could fail on
> persistent region as well due to a race.
> Geode by default recovering persistent region entry asynchronously, so there
> is a race that a get could land on an region entry not yet recovered from
> disk. And transaction will get a NOT_AVAILABLE value or fail with
> NotSerializableException: org.apache.geode.internal.cache.Token$NotAvailable
> if transaction initiated on a client.
> To avoid this race, user can set -Dgemfire.disk.recoverValuesSync system
> property to true to avoid this issue.
> This issue will be fixed in GEODE-1.7.0 along with fix of GEODE-5173.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
