jvarenina commented on a change in pull request #6036:
URL: https://github.com/apache/geode/pull/6036#discussion_r696449935
##########
File path:
geode-wan/src/distributedTest/java/org/apache/geode/internal/cache/wan/parallel/ParallelWANPropagationLoopBackDUnitTest.java
##########
@@ -365,9 +365,9 @@ public void
testParallelPropagationLoopBack3SitesNtoNTopologyPutFromOneDS() {
* Site-NY: dsid=1: senderId="ln": vm3, vm6
* NY site's sender's manual-start=true
*
- * Make sure the events are sent from LN to NY and will not be added into
tmpDroppedEvents
- * while normal events put from NY site can still be added to
tmpDroppedEvents
- * Start the sender, make sure the events in tmpDroppedEvents are sent to LN
finally
+ * Verify that events aren't added to tmpDroppedEvents after gateway-sender
is recovered in
+ * stopped state with manual-start=true, and that none of the events are
sent to remote site
+ * after sender is started.
Review comment:
Actually, in the beginning the gateway sender only recovers
ConcurrentParallelGatewaySenderEventProcessor in stopped state and creates
colocated gateway sender region (bucket are created later on in lazy manner
after data is received). Because ConcurrentParallelGatewaySenderEventProcessor
is recovered in stopped state the events are ignored and not stored in
temporary temporary queue.
Recovery of the queue region and
ConcurrentParallelGatewaySenderEventProcessor after main region is recovered:
```
vm5] [info 2021/08/26 11:04:39.784 CEST <RMI TCP
Connection(1)-192.168.1.104> tid=0x1b] Initializing region __PR
[vm5] [info 2021/08/26 11:04:39.789 CEST <RMI TCP
Connection(1)-192.168.1.104> tid=0x1b] Region __PR requesting initial image
from jakov(2368)<v1>:41005
[vm5] [info 2021/08/26 11:04:39.805 CEST <RMI TCP
Connection(1)-192.168.1.104> tid=0x1b] __PR is done getting image from
jakov(2368)<v1>:41005. isDeltaGII is false
[vm5] [info 2021/08/26 11:04:39.806 CEST <RMI TCP
Connection(1)-192.168.1.104> tid=0x1b] Initialization of region __PR completed
[vm5] [info 2021/08/26 11:04:39.821 CEST <RMI TCP
Connection(1)-192.168.1.104> tid=0x1b] Partitioned Region
/unstartedSenderShouldNotAddReceivedEventsIntoTmpDropped_PR is created with
prId=1
[vm5] [info 2021/08/26 11:04:39.866 CEST <RMI TCP
Connection(1)-192.168.1.104> tid=0x1b]
ConcurrentParallelGatewaySenderEventProcessor: dispatcher threads 1
[vm5] [info 2021/08/26 11:04:39.885 CEST <RMI TCP
Connection(1)-192.168.1.104> tid=0x1b] Partitioned Region
/ln_PARALLEL_GATEWAY_SENDER_QUEUE is created with prId=2
[vm5] [info 2021/08/26 11:04:39.904 CEST <RMI TCP
Connection(1)-192.168.1.104> tid=0x1b] Stopped
ParallelGatewaySender{id=ln,remoteDsId=2,isRunning =false}
```
Later on when data is received then the queue buckets are created together
with the main region buckets even thought manual start is set to true. This
wasn't the case before this impact. You could contact me on teams and we could
discuss this in more details
--
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]
