[
https://issues.apache.org/jira/browse/TINKERPOP3-903?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14974161#comment-14974161
]
Bob Briody commented on TINKERPOP3-903:
---------------------------------------
[~spmallette] is correct. The failures were based on the total payload of the
result, so adding/removing properties definitely altered the behavior. Small
results seem to always work fine. Some larger results would fail. Behavior was
also based on the how the buffer sizes were configured for Gremlin Server.
Lastly, failures were intermittent for the same result, although things could
be repeatable in isolation.
I'm definitely open to suggestions, but here's one possible plan... I will play
w/ 3.0.2 to find the "smallest" example graph that seems to fail consistently
and report back here so it can be used as a test case.
[~spmallette], what is {{writeBufferHighWaterMark}} set to when this test is
executed?
> Fix empty buffer return upon buffer capacity exceeded
> -----------------------------------------------------
>
> Key: TINKERPOP3-903
> URL: https://issues.apache.org/jira/browse/TINKERPOP3-903
> Project: TinkerPop 3
> Issue Type: Bug
> Components: driver, server
> Affects Versions: 3.0.2-incubating
> Reporter: stephen mallette
> Assignee: stephen mallette
> Fix For: 3.1.0-incubating
>
>
> Discussion: https://groups.google.com/forum/#!topic/gremlin-users/zWGAn-oS5PQ
> Return the underlying byte array output stream which will always contain the
> full serialized object.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
