DonalEvans commented on a change in pull request #7115:
URL: https://github.com/apache/geode/pull/7115#discussion_r751656704
##########
File path:
geode-for-redis/src/distributedTest/java/org/apache/geode/redis/internal/executor/string/StringsKillMultipleServersDUnitTest.java
##########
@@ -49,10 +50,19 @@
@BeforeClass
public static void classSetup() {
MemberVM locator = cluster.startLocatorVM(0);
- cluster.startRedisVM(1, locator.getPort());
- cluster.startRedisVM(2, locator.getPort());
- cluster.startRedisVM(3, locator.getPort());
- cluster.startRedisVM(4, locator.getPort());
+ int locatorPort = locator.getPort();
+ cluster.startRedisVM(1, x -> x
+ .withConnectionToLocator(locatorPort)
+ .withProperty(REDUNDANCY_ZONE, "A"));
+ cluster.startRedisVM(2, x -> x
+ .withConnectionToLocator(locatorPort)
+ .withProperty(REDUNDANCY_ZONE, "B"));
+ cluster.startRedisVM(3, x -> x
+ .withConnectionToLocator(locatorPort)
+ .withProperty(REDUNDANCY_ZONE, "A"));
+ cluster.startRedisVM(4, x -> x
+ .withConnectionToLocator(locatorPort)
+ .withProperty(REDUNDANCY_ZONE, "B"));
Review comment:
My understanding of redundancy zones may be wrong, but I think this
change doesn't necessarily guarantee that we lose all copies of a bucket when
crashing server3 and server4. If every secondary copy for server3 is on server2
and every secondary copy for server4 is on server1 then we can crash server3
and server4 but still have at least one copy of every bucket survive. I don't
know how easy it would be to do, but we could be sure of crashing the "right"
servers by finding which servers host the primary and the secondary for an
arbitrary bucket and crashing both.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
