geode-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jens Deppe (JIRA)" <j...@apache.org>
Subject [jira] [Created] (GEODE-520) GFSH Alter Region entry-time-to-live-expiration causes subsequent server restart to fail.
Date Fri, 30 Oct 2015 15:32:27 GMT
Jens Deppe created GEODE-520:
--------------------------------

             Summary: GFSH Alter Region entry-time-to-live-expiration causes subsequent server
restart to fail.
                 Key: GEODE-520
                 URL: https://issues.apache.org/jira/browse/GEODE-520
             Project: Geode
          Issue Type: Bug
          Components: management
            Reporter: Jens Deppe


h4. Summary
Example: alter region --name=/<Region-Name> --entry-time-to-live-expiration=<time
in seconds>
The above command execution causes unexpected behaviors in different cluster configurations
setup. Scenario-1 provides some options to keep the cluster running but scenario-2 makes the
cluster completely unstable.

h4. Scenario 1 (enable-cluster-configuration=false)
Restarting an exiting server or Adding a new server fails with new timeout value that was
already successfully applied to 'entry-time-to-live-expiration' value on a partitioned region
at runtime through gfsh.

Steps to Reproduce:-
# In cache.xml add below xml fragment inside {{<region-attributes>}} element for a partitioned
region.
{noformat}
<entry-time-to-live><expiration-attributes timeout="300" action="destroy"/></entry-time-to-live>
{noformat}
# Start the locator using {{--enable-cluster-configuration=false}}
# Start 2-3 servers using the cache.xml. Let's say server1, server2, server3
# Execute {{alter region --name=/<Region-Name> --entry-time-to-live-expiration=200}}
in gfsh and notice new value updated to all the running server.
{noformat}
gfsh>alter region --name=/Customer --entry-time-to-live-expiration=200
Member | Status
server1 | Region "/Customer" altered on "server1"
server2 | Region "/Customer" altered on "server2"
server3 | Region "/Customer" altered on "server3"
{noformat}
# Stop server2. Modify the cache.xml to have the new timeout value of 200 as the running servers
are already having new timeout value of 200 for the regions.
# Start the server2 and notice server start-up fails with below error.
{noformat}
Exception in thread "main" java.lang.IllegalStateException?: The entry time to live set in
RegionAttributes? is incompatible with entry time to live used by other distributed members.
at com.gemstone.gemfire.internal.cache.PartitionRegionConfigValidator?.validateExpirationAttributes(PartitionRegionConfigValidator?.java:198)
{noformat}
# Revert back the timeout to old value of 300. Start server2 and notice server2 will start
without any issue though timeout value is different that other cluster members.
# Adding a new server also behaves same as restarting an existing server described in Step-6
and Step-7



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message