You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@deltacloud.apache.org by "David Lutterkort (JIRA)" <ji...@apache.org> on 2011/07/14 18:46:59 UTC

[jira] [Resolved] (DTACLOUD-58) api/instance query performance with 8 vms in rhevm

     [ https://issues.apache.org/jira/browse/DTACLOUD-58?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

David Lutterkort resolved DTACLOUD-58.
--------------------------------------

    Resolution: Won't Fix

This isn't a Deltacloud issue; it's a combination of Aeolus polling all instances every 30s and RHEV-M 2.2 being slow to respond.

See
http://mail-archives.apache.org/mod_mbox/incubator-deltacloud-dev/201107.mbox/%3C20110714124958.GB2576@localhost.localdomain%3E

> api/instance query performance with 8 vms in rhevm
> --------------------------------------------------
>
>                 Key: DTACLOUD-58
>                 URL: https://issues.apache.org/jira/browse/DTACLOUD-58
>             Project: DeltaCloud
>          Issue Type: Bug
>         Environment: Aeolus-conductor configured for rhevm provider
> Rhevm running in Windows 2008 R2 Server with 1.5GB memory VM (running on a separate blade server)
> [root@hp-dl2x170g6-01 ~]# rpm -qa | egrep 'aeolus|deltacloud' | sort
> aeolus-all-0.3.0-0.el6.20110711131044git5bc7abf.noarch
> aeolus-conductor-0.3.0-0.el6.20110711131044git5bc7abf.noarch
> aeolus-conductor-daemons-0.3.0-0.el6.20110711131044git5bc7abf.noarch
> aeolus-conductor-doc-0.3.0-0.el6.20110711131044git5bc7abf.noarch
> aeolus-configure-2.0.1-0.el6.20110708134115gitab1e6dc.noarch
> condor-deltacloud-gahp-7.6.0-5dcloud.el6.x86_64
> deltacloud-core-0.3.9999-1308927004.el6.noarch
> libdeltacloud-0.9-1.el6.x86_64
> rubygem-aeolus-cli-0.0.1-1.el6.20110711131044git5bc7abf.noarch
> rubygem-deltacloud-client-0.3.1-1.el6.noarch
> [root@hp-dl2x170g6-01 ~]# 
>            Reporter: dave johnson
>            Assignee: David Lutterkort
>         Attachments: rhevm.perf.log
>
>
> I noticed that when I had over 7 rhev-h vm's running within rhevm, the CPU was running > 90% and the api console was practically scrolling continuously.  I looked at the deltacloud-core log file for rhevm and noticed that the query time of /api/instances were taking > 45 seconds to complete.  I started over adding a single vm every five minutes and noticed that there is a "heartbeat" from deltacloud-core every 30 seconds, if querying the API is taking longer than the heartbeat interval, deltacloud-core's heartbeart is causing performance issues with the rhevm api.  
> With 8 vm's running, deltacloud-core api/instabces query were timing out and throwing timeout exceptions.  
> # instances deployed through conductor
> # instance starts in deltacloud-core log
> 127.0.0.1 - - [13/Jul/2011 16:13:15] "POST /api/instances/7b33e17e-3563-4717-b8e4-e5d5f454a240/start HTTP/1.1" 204 - 2.3925
> 127.0.0.1 - - [13/Jul/2011 16:20:50] "POST /api/instances/b24424ae-ab1e-4974-ba97-c0e4a8a2a779/start HTTP/1.1" 204 - 2.0416
> 127.0.0.1 - - [13/Jul/2011 16:26:01] "POST /api/instances/20a5dcad-b053-4267-9e5b-46f40f9d42c7/start HTTP/1.1" 204 - 3.9677
> 127.0.0.1 - - [13/Jul/2011 16:31:48] "POST /api/instances/e79f7b0e-7cc8-4426-88d7-3a3e1728fbb6/start HTTP/1.1" 204 - 13.0901
> 127.0.0.1 - - [13/Jul/2011 16:35:55] "POST /api/instances/5327ddc2-8a95-43c9-8b00-80e28fa1a16b/start HTTP/1.1" 204 - 4.2185
> 127.0.0.1 - - [13/Jul/2011 16:41:23] "POST /api/instances/2d8624a7-ecd5-4e7a-a479-f2ae45f96a70/start HTTP/1.1" 204 - 7.6413
> 127.0.0.1 - - [13/Jul/2011 16:46:33] "POST /api/instances/0b67d735-cd54-4a12-ba43-d708b018e6d0/start HTTP/1.1" 204 - 6.6535
> 127.0.0.1 - - [13/Jul/2011 16:52:26] "POST /api/instances/d16a2597-0916-419f-83cd-86e41dc08ca6/start HTTP/1.1" 204 - 20.1581
> # api/instance load times climbing
> 127.0.0.1 - - [13/Jul/2011 16:15:22] "GET /api/instances HTTP/1.1" 200 1223 3.2926
> 127.0.0.1 - - [13/Jul/2011 16:20:12] "GET /api/instances HTTP/1.1" 200 1223 3.4131
> 127.0.0.1 - - [13/Jul/2011 16:25:17] "GET /api/instances HTTP/1.1" 200 2381 6.8360
> 127.0.0.1 - - [13/Jul/2011 16:30:54] "GET /api/instances HTTP/1.1" 200 3539 9.8815
> 127.0.0.1 - - [13/Jul/2011 16:35:17] "GET /api/instances HTTP/1.1" 200 4697 8.3200
> 127.0.0.1 - - [13/Jul/2011 16:40:51] "GET /api/instances HTTP/1.1" 200 5855 21.5512
> 127.0.0.1 - - [13/Jul/2011 16:45:00] "GET /api/instances HTTP/1.1" 200 7013 24.4277
> 127.0.0.1 - - [13/Jul/2011 16:50:05] "GET /api/instances HTTP/1.1" 200 8171 35.1951
> 127.0.0.1 - - [13/Jul/2011 16:50:07] "GET /api/instances HTTP/1.1" 200 8171 34.3853
> 127.0.0.1 - - [13/Jul/2011 16:50:14] "GET /api/instances HTTP/1.1" 200 8171 34.7330
> 127.0.0.1 - - [13/Jul/2011 16:50:19] "GET /api/instances HTTP/1.1" 200 8171 34.0917
> 127.0.0.1 - - [13/Jul/2011 16:50:32] "GET /api/instances HTTP/1.1" 200 8171 34.0356
> 127.0.0.1 - - [13/Jul/2011 16:50:57] "GET /api/instances HTTP/1.1" 200 8171 47.4070
> 127.0.0.1 - - [13/Jul/2011 16:51:05] "GET /api/instances HTTP/1.1" 200 8171 50.6648
> 127.0.0.1 - - [13/Jul/2011 16:51:08] "GET /api/instances HTTP/1.1" 200 8171 52.6800
> 127.0.0.1 - - [13/Jul/2011 16:51:09] "GET /api/instances HTTP/1.1" 200 8171 53.5951
> 127.0.0.1 - - [13/Jul/2011 16:51:29] "GET /api/instances HTTP/1.1" 200 8171 53.6821
> 127.0.0.1 - - [13/Jul/2011 16:51:34] "GET /api/instances HTTP/1.1" 200 8171 57.5881
> 127.0.0.1 - - [13/Jul/2011 16:51:35] "GET /api/instances HTTP/1.1" 200 8171 51.4359
> 127.0.0.1 - - [13/Jul/2011 16:51:38] "GET /api/instances HTTP/1.1" 200 8171 48.9456
> 127.0.0.1 - - [13/Jul/2011 16:51:42] "GET /api/instances HTTP/1.1" 200 8171 39.6616
> 127.0.0.1 - - [13/Jul/2011 16:52:01] "GET /api/instances HTTP/1.1" 200 9454 33.8931
> 127.0.0.1 - - [13/Jul/2011 16:52:13] "GET /api/instances HTTP/1.1" 200 9454 37.3036
> 127.0.0.1 - - [13/Jul/2011 16:52:20] "GET /api/instances HTTP/1.1" 200 9454 41.3128
> 127.0.0.1 - - [13/Jul/2011 16:52:58] "GET /api/instances HTTP/1.1" 200 9454 59.2965
> 127.0.0.1 - - [13/Jul/2011 16:53:04] "GET /api/instances HTTP/1.1" 500 190 60.0303
> 127.0.0.1 - - [13/Jul/2011 16:53:05] "GET /api/instances HTTP/1.1" 500 190 60.0688

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira