Bug 1493034

Summary:	Exclude etcd3 in OSCP 3.4 installation
Product:	OpenShift Container Platform	Reporter:	Vladislav Walek <vwalek>
Component:	Installer	Assignee:	Scott Dodson <sdodson>
Status:	CLOSED WONTFIX	QA Contact:	Johnny Liu <jialiu>
Severity:	high	Docs Contact:
Priority:	high
Version:	3.4.0	CC:	aos-bugs, gsapienz, jlee, jokerman, mmccomas, nnosenzo, sdodson, sgaikwad, vwalek
Target Milestone:	---	Keywords:	Reopened
Target Release:	3.4.z
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2018-04-10 15:51:35 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Vladislav Walek 2017-09-19 08:37:40 UTC

Description of problem:

When installing the OSCP 3.4, the etcd3 or etcd>3.0.x could be installed. This is not correct as the valid etcd version for OSCP 3.4 should be 3.0.X.
There should be some excluder or hard dependency for the packages as during the prod installation the incorrect version could be installed.
https://access.redhat.com/articles/2176281

Comment 1 Vladislav Walek 2017-09-19 08:40:27 UTC

Hi,

The etcd package is comming from rhel-7-server-extras-rpms, not from the ose package.

Comment 2 Scott Dodson 2017-09-19 13:48:14 UTC

Is there an actual problem encountered here? OCP 3.4 should work fine with etcd 3.0 or 3.1.

Comment 3 Scott Dodson 2017-09-19 13:52:26 UTC

Or 3.2.

Comment 4 Vladislav Walek 2017-09-19 14:02:50 UTC

Hello Scott,

based on https://access.redhat.com/articles/2176281 the 3.4 should be only 3.0. The 3.5 should be ok with 3.0 or 3.1.
Unfortunately, there is already etcd version 3.2 available from rhel-7-server-extras-rpms ?
Thx

Comment 5 Scott Dodson 2017-09-19 14:44:10 UTC

Then that document needs to be updated. Until an actual defect related to OCP 3.4 and etcd is observed this is NOTABUG.

Comment 7 Scott Dodson 2017-09-25 12:47:39 UTC

(In reply to jooho lee from comment #6)
> @Scott
> 
> Is there any chance upgrading etcd to etcd3 accidentally? If so, the scheme
> should be a problem.

No, you'd have to manually alter the master config to start using v3 storage. This should not be a problem.

Comment 10 Nicolas Nosenzo 2018-03-07 10:45:12 UTC

Re-opening this bug, since the upgrade from etcd3.1 to etcd3.2 break the etcd cluster. 

The reason for this, is that:

ETCD_CA_FILE is deprecated and replaced by ETCD_TRUSTED_CA_FILE
ETCD_PEER_CA_FILE is deprecated and replaced by ETCD_PEER_TRUSTED_CA_FILE

Reference:

https://coreos.com/etcd/docs/latest/v2/configuration.html#security-flags

So I'm wondering if we can avoid people to accidentally run a plain " yum update" and fall into this issue.

** More information about this upgrade problem in BZ [1529575](https://bugzilla.redhat.com/show_bug.cgi?id=1529575)

Comment 13 Scott Dodson 2018-04-10 15:51:35 UTC

https://bugzilla.redhat.com/show_bug.cgi?id=1529575 is a distinct bug related to etcd 3.2 options and etcd 3.1 configuration file and we're addressing that separately.