You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@trafodion.apache.org by "Selvaganesan Govindarajan (JIRA)" <ji...@apache.org> on 2017/07/26 02:54:00 UTC

[jira] [Created] (TRAFODION-2695) SSMP process ($ZSMxxx) sees too many opens from the master process

Selvaganesan Govindarajan created TRAFODION-2695:
----------------------------------------------------

             Summary: SSMP process ($ZSMxxx) sees too many opens from the master process
                 Key: TRAFODION-2695
                 URL: https://issues.apache.org/jira/browse/TRAFODION-2695
             Project: Apache Trafodion
          Issue Type: Improvement
            Reporter: Selvaganesan Govindarajan


The master processes like mxosrvr or sqlci opens SSMP process to send query started message and query finished messages. These messages are sent to provide the capability to cancel the query.  

I had observed the following with mxosrvr and mxssmp interactions:

Mxosrvr  opens a connection to mxssmp  
-	For get statistics command
        - Managed via ssmpManager_ in the context. This can have connections to all the ssmps in the cluster
-	To cancel a query
        -  Managed via cbServer_ in ExCancelTcb.  This connection is expected to go away when the cancel is passed on the mxssmp.
-	To Send query started /Query finished message
         - Managed via cbServer_ in ContextCli. If the cbServer_ is taken up already by the query with query started message pending, every statement being executed would create a connection to ssmp and managed via cbServer_ in ex_root_tcb of the query

I can see many (4) opens in mxssmp dump  from a mxosrvr, but I could account for only one connection on the mxosrvr core dump.  For some other clients, I have seen upto 6-8 opens. 

So, I would like to change into a common connection pool for ssmps. The common connection pool should be managed at the contextCli. 





--
This message was sent by Atlassian JIRA
(v6.4.14#64029)