[jira] [Updated] (TINKERPOP-2157) SparkStarBarrierInterceptor injects (Byte) 0
[
https://issues.apache.org/jira/browse/TINKERPOP-2157?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Daniel Kuppitz updated TINKERPOP-2157:
--
Summary: SparkStarBarrierInterceptor injects (Byte) 0 (was:
SparkStarBarrierInterceptor injects (Bte) 0)
> SparkStarBarrierInterceptor injects (Byte) 0
>
>
> Key: TINKERPOP-2157
> URL: https://issues.apache.org/jira/browse/TINKERPOP-2157
> Project: TinkerPop
> Issue Type: Bug
> Components: process
>Affects Versions: 3.4.1
>Reporter: Daniel Kuppitz
>Assignee: stephen mallette
>Priority: Major
>
> There are 3 tests that occasionally fail in
> {{SparkGraphComputerProcessIntegrateTest}}:
> * {{g_V_name_min}}
> * {{g_V_name_max}}
> * {{g_V_age_min}}
> In all cases, failure is a result of a mysterious {{(Byte) 0}}. Apparently,
> this zero gets injected by the {{SparkStarBarrierInterceptor}}. I came to
> this conclusion because I wasn't able to reproduce any of the failures after
> excluding this interceptor. However, the failures happen randomly*, so I
> can't say with 100% certainty that the interceptor is the problem.
>
> * on my system it never happened in the Docker builds, rarely in IntelliJ
> debug sessions, but quite often when I just run the test suite in IntelliJ
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Created] (TINKERPOP-2157) SparkStarBarrierInterceptor injects (Bte) 0
Daniel Kuppitz created TINKERPOP-2157:
-
Summary: SparkStarBarrierInterceptor injects (Bte) 0
Key: TINKERPOP-2157
URL: https://issues.apache.org/jira/browse/TINKERPOP-2157
Project: TinkerPop
Issue Type: Bug
Components: process
Affects Versions: 3.4.1
Reporter: Daniel Kuppitz
Assignee: stephen mallette
There are 3 tests that occasionally fail in
{{SparkGraphComputerProcessIntegrateTest}}:
* {{g_V_name_min}}
* {{g_V_name_max}}
* {{g_V_age_min}}
In all cases, failure is a result of a mysterious {{(Byte) 0}}. Apparently,
this zero gets injected by the {{SparkStarBarrierInterceptor}}. I came to this
conclusion because I wasn't able to reproduce any of the failures after
excluding this interceptor. However, the failures happen randomly*, so I can't
say with 100% certainty that the interceptor is the problem.
* on my system it never happened in the Docker builds, rarely in IntelliJ debug
sessions, but quite often when I just run the test suite in IntelliJ
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Commented] (TINKERPOP-2154) GraphBinary: Serializers should release resources in case of failures
[
https://issues.apache.org/jira/browse/TINKERPOP-2154?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16763512#comment-16763512
]
ASF GitHub Bot commented on TINKERPOP-2154:
---
jorgebay commented on pull request #1058: TINKERPOP-2154 GraphBinary: Release
working buffers on failure
URL: https://github.com/apache/tinkerpop/pull/1058
https://issues.apache.org/jira/browse/TINKERPOP-2154
Ensures that reference counts of allocated buffers are set to 0 when there
is a failure during serialization.
Includes a test for memory allocation when introducing failures.
VOTE +1
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> GraphBinary: Serializers should release resources in case of failures
> -
>
> Key: TINKERPOP-2154
> URL: https://issues.apache.org/jira/browse/TINKERPOP-2154
> Project: TinkerPop
> Issue Type: Bug
> Components: io
>Affects Versions: 3.4.0
>Reporter: Jorge Bay
>Assignee: Jorge Bay
>Priority: Major
> Fix For: 3.4.1
>
>
> {{TypeSerializer}} implementations should release the working {{ByteBuf}}
> instances when writing fails. This is specially concerning for composite
> types (List/Map/ByteCode/...) that can contain arbitrary nested types.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[GitHub] [tinkerpop] jorgebay opened pull request #1058: TINKERPOP-2154 GraphBinary: Release working buffers on failure
https://issues.apache.org/jira/browse/TINKERPOP-2154 Ensures that reference counts of allocated buffers are set to 0 when there is a failure during serialization. Includes a test for memory allocation when introducing failures. VOTE +1 [ Full content available at: https://github.com/apache/tinkerpop/pull/1058 ] This message was relayed via gitbox.apache.org for dev@tinkerpop.apache.org
