Bug 1600253

Summary: Out of memory when getting EC2 public images
Product: Red Hat CloudForms Management Engine Reporter: Matouš Mojžíš <mmojzis>
Component: ProvidersAssignee: Ladislav Smola <lsmola>
Status: CLOSED CURRENTRELEASE QA Contact: Matouš Mojžíš <mmojzis>
Severity: urgent Docs Contact:
Priority: high    
Version: 5.10.0CC: agrare, cpelland, gblomqui, jcheal, jfrey, jhardy, mmojzis, obarenbo, simaishi
Target Milestone: GAKeywords: Regression
Target Release: 5.10.0   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard: ec2:image:performance
Fixed In Version: 5.10.0.6 Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2019-02-11 13:59:48 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: Bug
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: CFME Core Target Upstream Version:
Embargoed:
Attachments:
Description Flags
evm.log errors none

Description Matouš Mojžíš 2018-07-11 20:39:54 UTC
Created attachment 1458219 [details]
evm.log errors

Description of problem:


Version-Release number of selected component (if applicable):
5.10.0.3

How reproducible:
Always

Steps to Reproduce:
1. Go to Configuration -> Advanced settings -> Set get_public_images to true
2. Add an ec2 provider
3.

Actual results:
Appliance stats being Out of memory after some time.
Even when I added some RAM(16G + 8G swap) it was out of memory.
In 5.9 I was able to get all the public images with default memory without issues. (Appliance was using only 4G of RAM)

Expected results:


Additional info:

Comment 3 Julian Cheal 2018-07-13 14:56:51 UTC
Hi Matouš

Do you have a 5.9 & a 5.10 appliance I can use to look at to compare the results to help me look at what could have changed.

Comment 7 Adam Grare 2018-07-17 15:36:43 UTC
This appears to be a general issue with graph refresh scanning, moving to core and assigning to Ladas

Comment 9 CFME Bot 2018-07-18 19:30:54 UTC
New commit detected on ManageIQ/manageiq/master:

https://github.com/ManageIQ/manageiq/commit/b89363252b51d9a28454429e6a2abceb9ab73f46
commit b89363252b51d9a28454429e6a2abceb9ab73f46
Author:     Ladislav Smola <lsmola>
AuthorDate: Wed Jul 18 10:01:02 2018 -0400
Commit:     Ladislav Smola <lsmola>
CommitDate: Wed Jul 18 10:01:02 2018 -0400

    Remove the dynaminc readers and writers

    Remove the dynaminc readers and writers since they can cause
    massive spikes in memory and processing time.

    Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1600253

 app/models/manager_refresh/inventory_object.rb | 46 -
 1 file changed, 46 deletions(-)

Comment 10 Matouš Mojžíš 2018-08-01 11:56:09 UTC
Verified in 5.10.0.6. It uses less memory now - whole cfme uses 6GB of memory after successfully refreshing EC2 provider with public images enabled.