[jira] [Commented] (TRAFODION-1516) An trafodion infrastructure program mxssmp dumps core at times

["Selvaganesan Govindarajan (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TRAFODION-1516?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14950322#comment-14950322 ] 

Selvaganesan Govindarajan commented on TRAFODION-1516:
------------------------------------------------------

The analysis of the core revealed the following:

A partially corrupted message was received by ssmp process and hence it aborted. The corrupted message is the request to get process statistics information. Upon analyzing the other end of the message, it was found it originated from DCS MASTER. The monitor map file was used to find out the program of the other end.

DCS MASTER sends this request when the query needs to be cancelled.   Looking at the message, it was found DCS Master was sending the request to find the recent qid of the process 63889. But while trafodion engine tries to parse this request, the pid become negative and this resulted in junk characters and hence ssmp died processing the request.

> An trafodion infrastructure program mxssmp dumps core at times
> --------------------------------------------------------------
>
>                 Key: TRAFODION-1516
>                 URL: https://issues.apache.org/jira/browse/TRAFODION-1516
>             Project: Apache Trafodion
>          Issue Type: Bug
>          Components: sql-exe
>    Affects Versions: 1.0 (pre-incubation)
>         Environment: This problem was observed in a customer environment
>            Reporter: Selvaganesan Govindarajan
>            Assignee: Selvaganesan Govindarajan
>             Fix For: 1.2-incubating
>
>   Original Estimate: 24h
>  Remaining Estimate: 24h
>
> Customer observed the core of mxssmp process at $MY_SQROOT/sql/scripts directory randomly.



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