[
https://issues.apache.org/jira/browse/IGNITE-7648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16369972#comment-16369972
]
Alexei Scherbakov commented on IGNITE-7648:
-------------------------------------------
[~gvvinblade],
The main point of the patch is to make grid more resilient to faulty
situations.
If we cannot connect to node via communication until failure detection
threshold is reached (it's the main point of patch to try harder), it's not
usable and can be safely removed from topology.
Look at [1] for details on how it's possible to come into situation then
discovery is not enough for node removal. This is the reason the feature was
enabled by default in first place.
It's also protects us from misbehaving nodes who pretend to be alive but do not
provide valid response to handshakes.
As for proposed change, I think it makes sense. It will allow us to stay within
failure detection time more precisely.
[1] https://issues.apache.org/jira/browse/IGNITE-4499
> Revert IGNITE_ENABLE_FORCIBLE_NODE_KILL system property.
> --------------------------------------------------------
>
> Key: IGNITE-7648
> URL: https://issues.apache.org/jira/browse/IGNITE-7648
> Project: Ignite
> Issue Type: Improvement
> Affects Versions: 2.3
> Reporter: Alexei Scherbakov
> Assignee: Alexei Scherbakov
> Priority: Major
> Fix For: 2.5
>
>
> IGNITE_ENABLE_FORCIBLE_NODE_KILL system property was introduced in
> IGNITE-5718 as a way to prevent unnecessary node drops in case of short
> network problems.
> I suppose it's wrong decision to fix it in such way.
> We had faced some issues in our production due to lack of automatic kicking
> of ill-behaving nodes (on example, hanging due to long GC pauses) until we
> realised the necessity of changing default behavior via property.
> Right solution is to kick nodes only if failure threshold is reached. Such
> behavior should be always enabled.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
