incubator-cloudstack-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sowmya Krishnan (JIRA)" <j...@apache.org>
Subject [jira] [Created] (CLOUDSTACK-1496) List API Performance: listAccounts failing with OOME for high values of pagesize (>1000 )
Date Mon, 04 Mar 2013 11:57:13 GMT
Sowmya Krishnan created CLOUDSTACK-1496:
-------------------------------------------

             Summary: List API Performance: listAccounts failing with OOME for high values
of pagesize (>1000 )
                 Key: CLOUDSTACK-1496
                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-1496
             Project: CloudStack
          Issue Type: Bug
      Security Level: Public (Anyone can view this level - this is the default.)
          Components: API
    Affects Versions: 4.1.0
         Environment: CentOS release 6.3 
            Reporter: Sowmya Krishnan
            Assignee: Min Chen
             Fix For: 4.1.0


listAccounts with higher values of pagesize is resulting in OOME and in general, showing low
performance for smaller page size values too.

Configuration for the performance set up is as follows:

(Using simulator set up)

(Advanced zone)
Accounts: 2117
Hosts: 1986
Users: 2116
Virtual machines: 3299

Server configurations:

Management Server :
=================
8 proc Intel(R) Xeon(R) CPU E5620  @ 2.40GHz processor
CentOS release 6.3 (Final)

Database: 
========
8 proc Intel(R) Xeon(R) CPU E5620  @ 2.40GHz processor
Red Hat Enterprise Linux Server release 6.2 (Santiago)
MySQL-server-5.5.21-1.linux2.6.x86_64 (InnoDB engine)

Following are the results for listAccounts for various pagesize values:
======================================================
(APIs are fired on 8080 as admin from a different client than the MS)

pagesize=10 : 0m30.944s
pagesize=50: 1m57.292s
pagesize=100: 2m59.700s
pagesize=500: 12m22.807s


In comparison to the results before API Optimization, for pagesize=100, the response was received
within 62 seconds before optimization whereas now, the result is 149 seconds

One difference though, is that the performance run before API optimization was done on a DB
server with much higher configuration (Quad core, 8 processor). But still, we shouldn't be
seeing OOME.

Following is the extract of the result for listAccounts&listAll=true (# of accounts: 2117)
...
 <title>Error 500 Java heap space</title> </head> <body><h2>HTTP
ERROR 500</h2> <p>Problem accessing /client/api. Reason: <pre> Java heap
space</pre></p><h3>Caused by:</h3><pre>java.lang.OutOfMemoryError:
Java heap space </pre> <hr /><i><small>Powered by Jetty://</small>

Maven Settings:
============
[root@centos63 ~]# echo $MAVEN_OPTS
-Xmx5120m -XX:MaxPermSize=1024m -Xdebug -Xrunjdwp:transport=dt_socket,address=8787,server=y,suspend=n



--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message