-
Notifications
You must be signed in to change notification settings - Fork 114
Cluster Inconsistent State Classification
Qi Zhou edited this page Apr 15, 2019
·
8 revisions
- Severity: CRITICAL
- Expected Cluster Behavior: Shutdown
- Possible Reason: Machine/Container crashes (and fdatasync of rocksdb is not properly implemented), cluster is incorrectly implemented
- Fixes:
- Manually remove the root block in the master (tools to be provided); or
- Automatically scan and detect the root block with missing minor blocks (tools to be provided)
- Severity: CRITICAL
- Expected Cluster Behavior: Shutdown
- Possible Reason: Machine/Container crashes (and fdatasync of rocksdb is not properly implemented), cluster is incorrectly implemented Fixes:
- Manually remove the minor block in the slave (tools to be provided); or
- Automatically scan and detect the minor block with missing root block (tools to be provided)
- Severity: WARNING
- Expected Cluster Behavior: Cluster should continue to work
- Possible Reason: Cluster shutdown before the minor block is committed
- Related fixes: https://github.com/QuarkChain/pyquarkchain/issues/487
- Severity: WARNING
- Expected Cluster Behavior: Cluster/Slave will fail to sync minor blocks
- Possible Reason: Cluster shutdown before the root block is committed
- Fixes:
- Manually add the root block to slaves (tools to be provided)
- Automatically scan and detect the root block in the master while it is missing in slaves (tools to be provided)
- Severity: CRITICAL
- Expected Cluster Behavior: Cluster will redo the block and broadcast the x-shard tx deposits again
- Possible Reason: Cluster shutdown before the minor block is committed
- Fix: https://github.com/QuarkChain/pyquarkchain/pull/493
Official Website: https://www.quarkchain.io/
Discord Developer Community: https://discord.me/quarkchain