Bug 1752871

Summary:	[RFE] support device-mapper dm-raid target (stuck) reshape reversal
Product:	[Fedora] Fedora	Reporter:	Heinz Mauelshagen <heinzm>
Component:	kernel	Assignee:	Heinz Mauelshagen <heinzm>
Status:	NEW ---	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	rawhide	CC:	agk, airlied, anprice, bmarzins, bmr, bskeggs, cfeist, extras-qa, hdegoede, heinzm, ichavero, itamar, jarodwilson, jbrassow, jeremy, jglisse, john.j5live, jonathan, josef, kernel-maint, kzak, linville, lvm-team, masami256, mchehab, mcsontos, mjg59, msnitzer, pasik, prajnoha, prockai, steved, zkabelac
Target Milestone:	---
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:	1752868	Environment:
Last Closed:		Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:	1752868
Bug Blocks:

Description Heinz Mauelshagen 2019-09-17 12:57:43 UTC

Description of problem:
There's no device-mapper dm-raid target support to reverse a long running, reshaping RaidLV conversion (e.g. an erroneously started raid5 with 5 -> 7 stripes conversion) during the process.  The conversion has to finish up first and only then reversal can be started.  This can consume a lot of time and resources on large storage or slow transports, thus being able to interrupt the ongoing reshape and revert it can avoid such costs.  For stuck reshapes caused by bad blocks hit during conversion, such reverting functionality is one way to bail out (or being able to replace bad SubLVs during the conversion which is impossible with lvm2 now).

Version-Release number of selected component (if applicable):
All so far

How reproducible:


Steps to Reproduce:
1. Fail component SubLV pair during reshape causing reshape to get stuck
2.
3.

Actual results:
Long runtime/resource consumption/getting stuck without lvm supported option to bail out.

Expected results:
Be able to interrupt ongoing reshape at any time (even when stuck) and reverse it.

Additional info:
Upstream kernel commit 2c810cddc44d6f95cef75df3f07fc0850ff92417 added reshape reversal support to MD