Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
-
None
Description
Currently, the code implementing sending and receiving of snapshots is already there. Some scenarios need to be implemented:
- Node that has data restarts successfully (trivial, but should make sure that 'startup snapshot' is 'applied' correctly)
- Start a cluster of 3 nodes, stop one of them, add some data, provoke the lagging node to restore itself from a snapshot, start it, make sure it catches up
- Test that a node fed with a snapshot can act as a leader later (including installing snapshots to other nodes)
- Test that a node that catches up using a snapshot gets updates happening concurrently with snapshot streaming, applied in the correct order
Attachments
Issue Links
- causes
-
IGNITE-18122 Track last applied term and group config in storages
- Resolved
- is blocked by
-
IGNITE-18421 Fix IncomingSnapshotCopier#cancel for the case of network problems
- Resolved
-
IGNITE-18495 Fix RAFT snapshot installation hang due to response swap on retry
- Resolved
- is related to
-
IGNITE-18170 Deadlock in TableManager#updateAssignmentInternal()
- Resolved
-
IGNITE-18428 After a RAFT snapshot install timed out, subsequent installs consistently failed
- Resolved
-
IGNITE-18432 SafeTimeCandidateManager#commitIndex() fails assertion if indices have different terms
- Resolved
-
IGNITE-18449 Select from a table fails after the table is created
- Resolved
-
IGNITE-18451 Node stop may hang if an index is not created yet
- Resolved
-
IGNITE-19233 Add integration test for adding new node to the cluster while installing RAFT snapshot
- Open
-
IGNITE-18430 Add integration test for a case when RAFT snapshot installation is cancelled in the middle
- Resolved
- relates to
-
IGNITE-18485 ItTableRaftSnapshotsTest. entriesKeepAppendedAfterSnapshotInstallation() sometimes hangs
- Resolved
- split from
-
IGNITE-17262 Implement RAFT snapshot streamer
- Resolved
- links to