[ 
https://issues.apache.org/jira/browse/RATIS-874?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Hanisha Koneru updated RATIS-874:
---------------------------------
    Description: 
This Jira aims to fix the following:
 # Before sending an appendEntry request to follower, leader checks the 
validity of the request be verifying if the follower has the previous log 
entry. But if the follower had installed a snapshot, the previous could be 
missing and the appendEntry would still be valid. Hence, the SnapshotIndex 
should be factored in while checking the validity of appendEntry request. 
Leader should store Follower's SnapshotIndex for this.
 # When follower receives appendEntry request, it checks the validity of log 
entry - the first index of the log entry is exactly 1 more than the last log 
index. During this check, the snapshotIndex should also be considered i.e. the 
first index of the log entry can be 1 more than the last log index or the 
snapshotIndex.
 # After Ratis server is restared, it loads all the available log segments. But 
logs with end index < last snapshot index should not be loaded. There can be 
gaps in the log segments upto the snapshotIndex if a snapshot was installed 
from leader node. 

  was:
This Jira aims to fix the following:
# Before sending an appendEntry request to follower, leader checks the validity 
of the request be verifying if the follower has the previous log entry. But if 
the follower had installed a snapshot, the previous could be missing and the 
appendEntry would still be valid. Hence, the SnapshotIndex should be factored 
in while checking the validity of appendEntry request. Leader should store 
Follower's SnapshotIndex for this.
# When follower receives appendEntry request, it checks the validity of log 
entry - the first index of the log entry is exactly 1 more than the last log 
index. During this check, the snapshotIndex should also be considered i.e. the 
first index of the log entry can be 1 more than the last log index or the 
snapshotIndex.



> Fix AppendEntry validity checks to take the SnapshotIndex into account
> ----------------------------------------------------------------------
>
>                 Key: RATIS-874
>                 URL: https://issues.apache.org/jira/browse/RATIS-874
>             Project: Ratis
>          Issue Type: Bug
>            Reporter: Hanisha Koneru
>            Assignee: Hanisha Koneru
>            Priority: Major
>         Attachments: RATIS-874.001.patch
>
>
> This Jira aims to fix the following:
>  # Before sending an appendEntry request to follower, leader checks the 
> validity of the request be verifying if the follower has the previous log 
> entry. But if the follower had installed a snapshot, the previous could be 
> missing and the appendEntry would still be valid. Hence, the SnapshotIndex 
> should be factored in while checking the validity of appendEntry request. 
> Leader should store Follower's SnapshotIndex for this.
>  # When follower receives appendEntry request, it checks the validity of log 
> entry - the first index of the log entry is exactly 1 more than the last log 
> index. During this check, the snapshotIndex should also be considered i.e. 
> the first index of the log entry can be 1 more than the last log index or the 
> snapshotIndex.
>  # After Ratis server is restared, it loads all the available log segments. 
> But logs with end index < last snapshot index should not be loaded. There can 
> be gaps in the log segments upto the snapshotIndex if a snapshot was 
> installed from leader node. 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to