[
https://issues.apache.org/jira/browse/IGNITE-19567?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Maksim Timonin updated IGNITE-19567:
------------------------------------
Description:
1.start a node with consistentId A;
2.insert some data;
3../control.sh --snapshot create full
4.kill A;
5.start a node with consistentId B;
6../control.sh --snapshot restore full.
this is OK.
But, if an incremental snapshot is created, an error will be reported during
recovery:
1.start a node with consistentId A;
2.insert some data;
3../control.sh --snapshot create full --incremental
4.kill A;
5.start a node with consistentId B;
6../control.sh --snapshot restore full --increment 1.
There is a wrong check. It verifies that increments match the full snapshot.
But instead of getting consistentId from the full snapshot metafile, it gets it
from the local node.
---
There is a wrong check. It verifies that increments match the full snapshot.
But instead of getting consistentId from the full snapshot metafile, it gets it
from the local node.
Possible solutions is:
# Calculate affinity assignment basing on existing snapshot data (partitions)
# Validate that this affinity is correct
# Check that local snapshot and incremental snapshots match
# Restore incremental snapshot in this case.
Can be implemented on SnapshotRestoreProcess#preload state.
was:
1.start a node with consistentId A;
2.insert some data;
3../control.sh --snapshot create full
4.kill A;
5.start a node with consistentId B;
6../control.sh --snapshot restore full.
this is OK.
But, if an incremental snapshot is created, an error will be reported during
recovery:
1.start a node with consistentId A;
2.insert some data;
3../control.sh --snapshot create full --incremental
4.kill A;
5.start a node with consistentId B;
6../control.sh --snapshot restore full --increment 1.
There is a wrong check. It verifies that increments match the full snapshot.
But instead of getting consistentId from the full snapshot metafile, it gets it
from the local node.
> Wrong consistentId is used for verifying incremental snapshot
> -------------------------------------------------------------
>
> Key: IGNITE-19567
> URL: https://issues.apache.org/jira/browse/IGNITE-19567
> Project: Ignite
> Issue Type: Bug
> Affects Versions: 2.15
> Reporter: Maksim Timonin
> Assignee: Maksim Timonin
> Priority: Major
> Labels: IEP-89
> Fix For: 2.16
>
>
> 1.start a node with consistentId A;
> 2.insert some data;
> 3../control.sh --snapshot create full
> 4.kill A;
> 5.start a node with consistentId B;
> 6../control.sh --snapshot restore full.
> this is OK.
> But, if an incremental snapshot is created, an error will be reported during
> recovery:
> 1.start a node with consistentId A;
> 2.insert some data;
> 3../control.sh --snapshot create full --incremental
> 4.kill A;
> 5.start a node with consistentId B;
> 6../control.sh --snapshot restore full --increment 1.
> There is a wrong check. It verifies that increments match the full snapshot.
> But instead of getting consistentId from the full snapshot metafile, it gets
> it from the local node.
>
> ---
> There is a wrong check. It verifies that increments match the full snapshot.
> But instead of getting consistentId from the full snapshot metafile, it gets
> it from the local node.
>
> Possible solutions is:
> # Calculate affinity assignment basing on existing snapshot data (partitions)
> # Validate that this affinity is correct
> # Check that local snapshot and incremental snapshots match
> # Restore incremental snapshot in this case.
> Can be implemented on SnapshotRestoreProcess#preload state.
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
