[JBoss JIRA] (ISPN-11304) Allow scaling up without state transfer
by Tristan Tarrant (Jira)
[ https://issues.redhat.com/browse/ISPN-11304?page=com.atlassian.jira.plugi... ]
Tristan Tarrant updated ISPN-11304:
-----------------------------------
Status: Resolved (was: Pull Request Sent)
Fix Version/s: 12.0.0.Final
(was: 11.0.2.Final)
Resolution: Done
> Allow scaling up without state transfer
> ---------------------------------------
>
> Key: ISPN-11304
> URL: https://issues.redhat.com/browse/ISPN-11304
> Project: Infinispan
> Issue Type: Feature Request
> Components: Core
> Affects Versions: 10.1.1.Final
> Reporter: Dan Berindei
> Assignee: Dan Berindei
> Priority: Major
> Fix For: 12.0.0.Final
>
>
> We should allow a cache to scale up without performing any state transfer, but without losing the data.
> To simplify things, the initial version will support a single owner, and will assume that only one node is being added at a time.
> The cache must be accessible remotely, but since information about the location of the keys is not accessible from the client, the client is expected to ignore ownership and use a round-robin access strategy.
--
This message was sent by Atlassian Jira
(v7.13.8#713008)
4 years, 5 months
[JBoss JIRA] (ISPN-11981) Stale immortal entry metadata
by Tristan Tarrant (Jira)
[ https://issues.redhat.com/browse/ISPN-11981?page=com.atlassian.jira.plugi... ]
Tristan Tarrant updated ISPN-11981:
-----------------------------------
Status: Resolved (was: Pull Request Sent)
Fix Version/s: 12.0.0.Dev01
(was: 11.0.2.Final)
Resolution: Done
> Stale immortal entry metadata
> -----------------------------
>
> Key: ISPN-11981
> URL: https://issues.redhat.com/browse/ISPN-11981
> Project: Infinispan
> Issue Type: Bug
> Components: Core
> Affects Versions: 11.0.0.CR1, 10.1.8.Final
> Reporter: Dan Berindei
> Assignee: Dan Berindei
> Priority: Major
> Fix For: 12.0.0.Dev01
>
>
> When an entry is updated, {{InternalEntryFactoryImpl}} tries to reuse the existing {{InternalCacheEntry}} instance instead of creating a new one. The condition is that both the new and the old entry have the same implementation implementation: immortal, immortal w/ external metadata, transient, transient w/ metadata etc.
> Because {{MetadataImmortalCacheEntry}} extends {{ImmortalCacheEntry}}, that logic is broken: when an immortal entry with external metadata is replaced with a new entry without external metadata, the external metadata is not removed.
> This impacts the anchored keys module, which I am changing to store the key location in the metadata (currently it is stored as a value, and it is broken with BINARY/OFF_HEAP storage). When a node leaves and an entry located on the leaver is updated, the key is written on the newest joiner. The joiner already has a local entry with {{RemoteMetadata}}, and is trying to write a value without any metadata.
--
This message was sent by Atlassian Jira
(v7.13.8#713008)
4 years, 5 months
[JBoss JIRA] (ISPN-11304) Allow scaling up without state transfer
by Tristan Tarrant (Jira)
[ https://issues.redhat.com/browse/ISPN-11304?page=com.atlassian.jira.plugi... ]
Tristan Tarrant updated ISPN-11304:
-----------------------------------
Fix Version/s: 12.0.0.Dev01
(was: 12.0.0.Final)
> Allow scaling up without state transfer
> ---------------------------------------
>
> Key: ISPN-11304
> URL: https://issues.redhat.com/browse/ISPN-11304
> Project: Infinispan
> Issue Type: Feature Request
> Components: Core
> Affects Versions: 10.1.1.Final
> Reporter: Dan Berindei
> Assignee: Dan Berindei
> Priority: Major
> Fix For: 12.0.0.Dev01
>
>
> We should allow a cache to scale up without performing any state transfer, but without losing the data.
> To simplify things, the initial version will support a single owner, and will assume that only one node is being added at a time.
> The cache must be accessible remotely, but since information about the location of the keys is not accessible from the client, the client is expected to ignore ownership and use a round-robin access strategy.
--
This message was sent by Atlassian Jira
(v7.13.8#713008)
4 years, 5 months
[JBoss JIRA] (ISPN-11997) Remove jackson databind
by Tristan Tarrant (Jira)
[ https://issues.redhat.com/browse/ISPN-11997?page=com.atlassian.jira.plugi... ]
Tristan Tarrant updated ISPN-11997:
-----------------------------------
Status: Resolved (was: Pull Request Sent)
Resolution: Done
> Remove jackson databind
> ------------------------
>
> Key: ISPN-11997
> URL: https://issues.redhat.com/browse/ISPN-11997
> Project: Infinispan
> Issue Type: Enhancement
> Components: REST
> Affects Versions: 11.0.0.Final
> Reporter: Gustavo Fernandes
> Assignee: Gustavo Fernandes
> Priority: Major
> Fix For: 12.0.0.Final
>
>
> Jackson databind uses serialization and the native server needs rules to support it. Also it's desirable to reduce the number of dependencies in the REST server to avoid image respins due to CVEs
--
This message was sent by Atlassian Jira
(v7.13.8#713008)
4 years, 5 months
[JBoss JIRA] (ISPN-12139) Cache.size() always returns 0 for invalidation caches
by Paul Ferraro (Jira)
[ https://issues.redhat.com/browse/ISPN-12139?page=com.atlassian.jira.plugi... ]
Paul Ferraro updated ISPN-12139:
--------------------------------
Description:
It seems that Cache.size() always returns 0 for an invalidation cache.
See attached reproducer.
Filing as a blocker as this is a regression since 10.1.x.
was:
It seems that Cache.size() always returns 0 for an invalidation cache.
See attached reproducer.
> Cache.size() always returns 0 for invalidation caches
> -----------------------------------------------------
>
> Key: ISPN-12139
> URL: https://issues.redhat.com/browse/ISPN-12139
> Project: Infinispan
> Issue Type: Bug
> Components: Core
> Affects Versions: 11.0.1.Final
> Reporter: Paul Ferraro
> Priority: Blocker
> Attachments: Test.java
>
>
> It seems that Cache.size() always returns 0 for an invalidation cache.
> See attached reproducer.
> Filing as a blocker as this is a regression since 10.1.x.
--
This message was sent by Atlassian Jira
(v7.13.8#713008)
4 years, 5 months