-
Bug
-
Resolution: Done
-
Major
-
10.1.0.Final
-
None
-
DataGrid Sprint #36, DataGrid Sprint #37, DataGrid Sprint #38, DataGrid Sprint #39
When a cache is shutting down, the topology id is increased to Integer.MAX_VALUE.
If a component uses StateTransferLock.topologyUpdate() to retry in the next topology and doesn't explicitly check if the cache is running, it could retry 2^^31 times before giving up.
This causes StreamDistPartitionHandlingTest.clearContent to hang randomly:
11:58:14,655 TRACE (testng-Test:[]) [StateTransferLockImpl] Signalling topology 2147483647 is installed
11:58:14,655 TRACE (testng-Test:[]) [ClusterPublisherManagerImpl] Segments {0-3 8 12 21-22 25-26 29 31 35 37-39 41-42 47 52-54 56-58 73-75 77 85 93-95 97-99 104-105 107 118 121-122 126-129 133 140 145 149 151-154 161 167-168 170-171 177-178 186 189-190 195-196 199-205 212-214 216-217 221-222 224 226-227 232-233 243-244} not completed - retrying
11:58:14,655 TRACE (testng-Test:[]) [ClusterPublisherManagerImpl] Retrying segments {0-3 8 12 21-22 25-26 29 31 35 37-39 41-42 47 52-54 56-58 73-75 77 85 93-95 97-99 104-105 107 118 121-122 126-129 133 140 145 149 151-154 161 167-168 170-171 177-178 186 189-190 195-196 199-205 212-214 216-217 221-222 224 226-227 232-233 243-244} after 16 is installed for Test-NodeA-12596#7046
...
12:03:16,127 TRACE (testng-Test:[]) [ClusterPublisherManagerImpl] Segments {0-3 8 12 21-22 25-26 29 31 35 37-39 41-42 47 52-54 56-58 73-75 77 85 93-95 97-99 104-105 107 118 121-122 126-129 133 140 145 149 151-154 161 167-168 170-171 177-178 186 189-190 195-196 199-205 212-214 216-217 221-222 224 226-227 232-233 243-244} not completed - retrying
12:03:16,127 TRACE (testng-Test:[]) [ClusterPublisherManagerImpl] Retrying segments {0-3 8 12 21-22 25-26 29 31 35 37-39 41-42 47 52-54 56-58 73-75 77 85 93-95 97-99 104-105 107 118 121-122 126-129 133 140 145 149 151-154 161 167-168 170-171 177-178 186 189-190 195-196 199-205 212-214 216-217 221-222 224 226-227 232-233 243-244} after 16 is installed for Test-NodeA-12596#7046
...