ignite-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Artem Shutak (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (IGNITE-608) [Test] Fix cache with portableEnabled=true
Date Thu, 26 Mar 2015 12:47:56 GMT

     [ https://issues.apache.org/jira/browse/IGNITE-608?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Artem Shutak updated IGNITE-608:
--------------------------------
    Description: 
1. Portable object must be deserialized before return to user (see cache 'get' implementation),
looks like now it is not implemented for put/remove return value, peek, reload (need check
all cache public API).
2. Need to make sure that all tests in GridCachePortableEnabledSelfTestSuite pass (add suite
to TC after that).

Comment from GG-8743:
- Fixed all found issues with portables except metadata update inside explicit transaction.
If metadata is updated then method GridEntPortableProcessor.addMeta is called which can execute
'invoke' on system cache. For now just changed it to execute invoke in another thread (using
'CU.outTx'). Is it possible to add internal method to execute 'invoke' without enlisting system
cache in transaction? If this is not easy to add I think we can leave it as is since metadata
update is not expecte to happen frequently.

See GG-9271 and GG-8743. Also see TODOs in java-code for this issue.

  was:
1. Portable object must be deserialized before return to user (see cache 'get' implementation),
looks like now it is not implemented for put/remove return value, peek, reload (need check
all cache public API).
2. Need to make sure that all tests in GridCachePortableEnabledSelfTestSuite pass (add suite
to TC after that).

Comment from GG-8743:
- Fixed all found issues with portables except metadata update inside explicit transaction.
If metadata is updated then method GridEntPortableProcessor.addMeta is called which can execute
'invoke' on system cache. For now just changed it to execute invoke in another thread (using
'CU.outTx'). Is it possible to add internal method to execute 'invoke' without enlisting system
cache in transaction? If this is not easy to add I think we can leave it as is since metadata
update is not expecte to happen frequently.

See GG-9271 and GG-8743. Also see TODOs in java-code for this issue


> [Test] Fix cache with portableEnabled=true
> ------------------------------------------
>
>                 Key: IGNITE-608
>                 URL: https://issues.apache.org/jira/browse/IGNITE-608
>             Project: Ignite
>          Issue Type: Bug
>            Reporter: Artem Shutak
>            Assignee: Yakov Zhdanov
>              Labels: failedtest
>             Fix For: sprint-3
>
>
> 1. Portable object must be deserialized before return to user (see cache 'get' implementation),
looks like now it is not implemented for put/remove return value, peek, reload (need check
all cache public API).
> 2. Need to make sure that all tests in GridCachePortableEnabledSelfTestSuite pass (add
suite to TC after that).
> Comment from GG-8743:
> - Fixed all found issues with portables except metadata update inside explicit transaction.
If metadata is updated then method GridEntPortableProcessor.addMeta is called which can execute
'invoke' on system cache. For now just changed it to execute invoke in another thread (using
'CU.outTx'). Is it possible to add internal method to execute 'invoke' without enlisting system
cache in transaction? If this is not easy to add I think we can leave it as is since metadata
update is not expecte to happen frequently.
> See GG-9271 and GG-8743. Also see TODOs in java-code for this issue.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message