Dmitry Sherstobitov created IGNITE-8874:
-------------------------------------------
Summary: Blinking node in cluster may cause data corruption
Key: IGNITE-8874
URL: https://issues.apache.org/jira/browse/IGNITE-8874
Project: Ignite
Issue Type: Bug
Affects Versions: 2.5
Reporter: Dmitry Sherstobitov
All caches with 2 backups
4 nodes in cluster
# Start cluster, load data
# Start transactional loading (8 threads, 100 ops/second put/get in each op)
# Repeat 10 times: kill one node, clean LFS, start node again, wait for
rebalance
# Check idle_verify, check data corruption
Here is idle_verify report:
node2 - node that was blinking while test. Update counter are equal between
partitions but data is different.
{code:java}
Conflict partition: PartitionKey [grpId=374280886, grpName=cache_group_3,
partId=41]
Partition instances: [PartitionHashRecord [isPrimary=true, partHash=885018783,
updateCntr=16, size=15, consistentId=node4], PartitionHashRecord
[isPrimary=false, partHash=885018783, updateCntr=16, size=15,
consistentId=node3], PartitionHashRecord [isPrimary=false, partHash=-357162793,
updateCntr=16, size=15, consistentId=node2]]
Conflict partition: PartitionKey [grpId=1586135625, grpName=cache_group_1_015,
partId=15]
Partition instances: [PartitionHashRecord [isPrimary=true, partHash=-562597978,
updateCntr=22, size=16, consistentId=node3], PartitionHashRecord
[isPrimary=false, partHash=-562597978, updateCntr=22, size=16,
consistentId=node1], PartitionHashRecord [isPrimary=false, partHash=780813725,
updateCntr=22, size=16, consistentId=node2]]
Conflict partition: PartitionKey [grpId=374280885, grpName=cache_group_2,
partId=75]
Partition instances: [PartitionHashRecord [isPrimary=true,
partHash=-1500797699, updateCntr=21, size=16, consistentId=node3],
PartitionHashRecord [isPrimary=false, partHash=-1500797699, updateCntr=21,
size=16, consistentId=node1], PartitionHashRecord [isPrimary=false,
partHash=-1592034435, updateCntr=21, size=16, consistentId=node2]]
Conflict partition: PartitionKey [grpId=374280884, grpName=cache_group_1,
partId=713]
Partition instances: [PartitionHashRecord [isPrimary=false, partHash=-63058826,
updateCntr=4, size=2, consistentId=node3], PartitionHashRecord [isPrimary=true,
partHash=-63058826, updateCntr=4, size=2, consistentId=node1],
PartitionHashRecord [isPrimary=false, partHash=670869467, updateCntr=4, size=2,
consistentId=node2]]
Conflict partition: PartitionKey [grpId=374280886, grpName=cache_group_3,
partId=11]
Partition instances: [PartitionHashRecord [isPrimary=false,
partHash=-224572810, updateCntr=17, size=16, consistentId=node3],
PartitionHashRecord [isPrimary=true, partHash=-224572810, updateCntr=17,
size=16, consistentId=node1], PartitionHashRecord [isPrimary=false,
partHash=176419075, updateCntr=17, size=16, consistentId=node2]]{code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
