[jboss-jira] [JBoss JIRA] (WFWIP-311) OpenShift XP image doubled its size to 2GB compared to 7.3.0.GA 1GB

[Jeff Mesnil (Jira)]

    [ https://issues.redhat.com/browse/WFWIP-311?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14053354#comment-14053354 ] 

Jeff Mesnil commented on WFWIP-311:
-----------------------------------

I don't think it explains the difference in size but the comparison has too many variables
* The XP image is on JDK8 / RHEL7
* the EAP 7.3 and CD19 images are on JDK11 / RHEL8
* The XP image will be bigger than the EAP 7.3 images as it contains additional bits for the MicroProfile extensions (but this will not double its size).

I don't think there is anything specifically relevant to the XP images. It seems more related to the 7.3.1 codebase / image generation.
Once we have test builds for EAP 7.3.1 images, we will be able to locate more accurately the source of this size increase and make sure we address it.

> OpenShift XP image doubled its size to 2GB compared to 7.3.0.GA 1GB
> -------------------------------------------------------------------
>
>                 Key: WFWIP-311
>                 URL: https://issues.redhat.com/browse/WFWIP-311
>             Project: WildFly WIP
>          Issue Type: Bug
>          Components: OpenShift
>            Reporter: Martin Choma
>            Assignee: Jeff Mesnil
>            Priority: Major
>         Attachments: Screenshot from 2020-04-23 06-56-37.png
>
>
> I have started to play with OpenShift XP image {{docker-registry.upshift.redhat.com/kwills/eap-xp1-openjdk8-openshift-rhel7:EAP7-1484}} and I have noticed it is nearly 2GB big. Used to be 0.92 GB for 7.3.0.GA or CD19.
> Also strange is there is much more layers compared to several layers in previous releases
> From dive output it seems every bash command result in own layer. Dive is reporting 624MB of waste space. 
>  !Screenshot from 2020-04-23 06-56-37.png|thumbnail! 
> I am setting priority just to major because it is not official image yet. But if this will be the case of official image as well, I will raise priority. In OCP 4 there is no image/layer cache so image must be "downloaded" for each build from registry into pod.
> [1]
> {code}
> docker-registry.upshift.redhat.com/kwills/eap-xp1-openjdk8-openshift-rhel7:EAP7-1484			1.93 GB
> registry-proxy.engineering.redhat.com/rh-osbs/jboss-eap-7-eap73-openjdk11-openshift-rhel8:latest	920 MB
> registry-proxy.engineering.redhat.com/rh-osbs/jboss-eap-7-tech-preview-eap-cd-openshift-rhel8:19.0-2    921 MB
> {code}



--
This message was sent by Atlassian Jira
(v7.13.8#713008)