You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Anton Vinogradov (JIRA)" <ji...@apache.org> on 2017/11/23 12:06:01 UTC
[jira] [Updated] (IGNITE-7005) Ability to disable WAL (Recoverable
case)
[ https://issues.apache.org/jira/browse/IGNITE-7005?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Anton Vinogradov updated IGNITE-7005:
-------------------------------------
Description:
In addition to non recoverable case(IGNITE-7003):
On WAL disabling we should (on each node)
- trigger exchange to guarantie consistent state
- schedule new checkpoint. This checkpoint should be recorded to special place (temporary checkpoint location), to prevent damage of latest one.
All new checkpoints should update temporary checkpoint.
On WAL enabling we should (on each node) after all nodes reported that checkpoints finished
- merge temp checkpoint with stable (scheduled before WAL disabling)
- clean WAL
before enabling proxies
On any failure during loading (while WAL disabled or enabling) we should be able to reactivate cluster with
- data from original checkpoints & WAL for affected caches
- latest state for non affected caches
Failover:
Any topology change should be covered(while WAL disabled or enabling)
- Node(s) Left (inc. coordinator)
- Node(s) Join
was:
In addition to non recoverable case:
On WAL disabling we should (on each node)
- trigger exchange to guarantie consistent state
- schedule new checkpoint. This checkpoint should be recorded to special place (temporary checkpoint location), to prevent damage of latest one.
All new checkpoints should update temporary checkpoint.
On WAL enabling we should (on each node) after all nodes reported that checkpoints finished
- merge temp checkpoint with stable (scheduled before WAL disabling)
- clean WAL
before enabling proxies
On any failure during loading (while WAL disabled or enabling) we should be able to reactivate cluster with
- data from original checkpoints & WAL for affected caches
- latest state for non affected caches
Failover:
Any topology change should be covered(while WAL disabled or enabling)
- Node(s) Left (inc. coordinator)
- Node(s) Join
> Ability to disable WAL (Recoverable case)
> -----------------------------------------
>
> Key: IGNITE-7005
> URL: https://issues.apache.org/jira/browse/IGNITE-7005
> Project: Ignite
> Issue Type: Sub-task
> Components: persistence
> Reporter: Anton Vinogradov
> Assignee: Anton Vinogradov
>
> In addition to non recoverable case(IGNITE-7003):
> On WAL disabling we should (on each node)
> - trigger exchange to guarantie consistent state
> - schedule new checkpoint. This checkpoint should be recorded to special place (temporary checkpoint location), to prevent damage of latest one.
> All new checkpoints should update temporary checkpoint.
> On WAL enabling we should (on each node) after all nodes reported that checkpoints finished
> - merge temp checkpoint with stable (scheduled before WAL disabling)
> - clean WAL
> before enabling proxies
> On any failure during loading (while WAL disabled or enabling) we should be able to reactivate cluster with
> - data from original checkpoints & WAL for affected caches
> - latest state for non affected caches
> Failover:
> Any topology change should be covered(while WAL disabled or enabling)
> - Node(s) Left (inc. coordinator)
> - Node(s) Join
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)