[ 
https://issues.apache.org/jira/browse/TRAFODION-953?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14720943#comment-14720943
 ] 

Suresh Subbiah commented on TRAFODION-953:
------------------------------------------

Anuradha (anuradha-hegde) on 2015-01-28
Changed in trafodion:
assignee:       nobody → Tharak Capirala (capirala-tharaknath)
milestone:      none → r1.0.1
Arvind Narain (arvind-narain) wrote on 2015-03-05:      #1
Partial fix put in stable 1.0.1 and master via fixes in connectivity and sql. 
Will update with email chains

Changed in trafodion:
assignee:       Tharak Capirala (capirala-tharaknath) → Arvind Narain 
(arvind-narain)
status: New → In Progress
milestone:      r1.0.1 → r1.1
Arvind Narain (arvind-narain) wrote on 2015-03-09:      #2
Along with this change https://review.trafodion.org/#/c/1189/ following fixes 
went into stable 1.0 and in mainline to handle memory leaks observed. Closing 
this bug as fix committed. Will reopen if leaks are observed in mxosrvr code.

1) memory leak and salted partitions set up fix -
https://review.trafodion.org/#/c/1106/

2) sqlplan disable - more for perfomance issues noticed.
https://review.trafodion.org/#/c/1103/
https://review.trafodion.org/#/c/1104/
=======================================

Some more fixes in mainline - including
1) Fix to take care of internal statement leak
https://review.trafodion.org/#/c/1180/

Changed in trafodion:
status: In Progress → Fix Committed


> LP Bug: 1415165 - MXOSRVR memory leak when stats are enabled
> ------------------------------------------------------------
>
>                 Key: TRAFODION-953
>                 URL: https://issues.apache.org/jira/browse/TRAFODION-953
>             Project: Apache Trafodion
>          Issue Type: Bug
>          Components: connectivity-mxosrvr
>            Reporter: Guy Groulx
>            Assignee: Arvind Narain
>            Priority: Blocker
>             Fix For: 1.1 (pre-incubation)
>
>
> Running longevity testing.
> On one system ran for 72 hours with stats disabled.    We did not see system 
> memory issue on the system.
> After 24 hr on another system with stats enabled, we're now very low on 
> memory.
> /home/squser2/guy/memleak> ps v -C mxosrvr
>   PID TTY      STAT   TIME  MAJFL   TRS   DRS   RSS %MEM COMMAND
> 42988 ?        Sl   179:52      0  1113 2187278 1139452  0.8 mxosrvr -ZKHOST 
> n004.cm.cluster:2181,n005.cm.cluster:2181,n006.cm.clust
> 42999 ?        Sl   194:24      0  1113 2186534 1128488  0.8 mxosrvr -ZKHOST 
> n004.cm.cluster:2181,n005.cm.cluster:2181,n006.cm.clust
> 43226 ?        Sl     1:17      0  1113 372110 21164  0.0 mxosrvr -ZKHOST 
> n004.cm.cluster:2181,n005.cm.cluster:2181,n006.cm.cluster:
> 43359 ?        Sl     1:16      0  1113 372110 21160  0.0 mxosrvr -ZKHOST 
> n004.cm.cluster:2181,n005.cm.cluster:2181,n006.cm.cluster:
> 43404 ?        Sl     1:17      0  1113 372110 21152  0.0 mxosrvr -ZKHOST 
> n004.cm.cluster:2181,n005.cm.cluster:2181,n006.cm.cluster:
> 43406 ?        Sl     1:16      0  1113 372110 21160  0.0 mxosrvr -ZKHOST 
> n004.cm.cluster:2181,n005.cm.cluster:2181,n006.cm.cluster:
> 43462 ?        Sl     1:18      0  1113 372110 21152  0.0 mxosrvr -ZKHOST 
> n004.cm.cluster:2181,n005.cm.cluster:2181,n006.cm.cluster:
> 43534 ?        Sl   189:02      0  1113 2170298 1120896  0.8 mxosrvr -ZKHOST 
> n004.cm.cluster:2181,n005.cm.cluster:2181,n006.cm.clust
> See how the mxosrvr in use are using over 1GB of resident space.
> The ones now being used are still at 21mb only.



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

Reply via email to