Bug 1600058 – VDO can go read-only, lose sparsely-written data when parts are discarded. [rhel-7.5.z]

Bugzilla will be upgraded to version 5.0. The upgrade date is tentatively scheduled for 2 December 2018, pending final testing and feedback.

Bug 1600058 - VDO can go read-only, lose sparsely-written data when parts are discarded. [rhel-7.5.z]

Summary:	VDO can go read-only, lose sparsely-written data when parts are discarded. [r...

Status:	CLOSED ERRATA

Aliases:	None

Product:	Red Hat Enterprise Linux 7
Classification:	Red Hat
Component:	kmod-kvdo (Show other bugs)
Sub Component:
Version:	7.6
Hardware:	Unspecified Unspecified

Priority	high Severity unspecified
Target Milestone:	rc
Target Release:	---
Assigned To:	vdo-internal
QA Contact:	Jakub Krysl
Docs Contact:	Marek Suchanek

URL:
Whiteboard:
Keywords:	ZStream

Depends On:	1589249
Blocks:
	Show dependency tree / graph

Reported:	2018-07-11 06:31 EDT by Jaroslav Reznik
Modified:	2018-09-24 07:23 EDT (History)
CC List:	6 users (show)

See Also:
Fixed In Version:	6.1.0.179
Doc Type:	If docs needed, set a value
Doc Text:	Previously, VDO did not correctly reinitialize certain structures when a discard spanned logical addresses on two different block map trees. As a consequence, a discard operation sometimes switched the affected VDO volume to read-only mode or corrupted data on it in rare cases. With this update, VDO now reinitializes the structures correctly, and the described problem no longer occurs.
Story Points:	---
Clone Of:	1589249
Environment:
Last Closed:	2018-08-16 10:19:04 EDT
Type:	---
Regression:	---
Mount Type:	---
Documentation:	---
CRM:
Verified Versions:
Category:	---
oVirt Team:	---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---

Attachments	(Terms of Use)
Add an attachment (proposed patch, testcase, etc.)

External Trackers
Tracker	ID	Priority	Status	Summary	Last Updated
Red Hat Product Errata	RHBA-2018:2450	None	None	None	2018-08-16 10:19 EDT

Groups: None (edit)

Description Jaroslav Reznik 2018-07-11 06:31:05 EDT

This bug has been copied from bug #1589249 and has been proposed to be backported to 7.5 z-stream (EUS).

Comment 3 Jakub Krysl 2018-07-24 09:53:36 EDT

tested with kvdo 6.1.0.181, could not hit the issue after 50 iterations. Original reproducer was able to hit it after 8.

Comment 5 Jakub Krysl 2018-08-09 10:29:15 EDT

Hi Marek,

Yes, it is correct.
Is seems a bit more serious than it really is, but I have no idea how to incorporate all the needed information and keep it short at the same time.
VDO max_discard_size is really small so a discard spanning 2 trees (one of which is not fully allocated) has really small chance of happening on default VDO device. Though it is possible to increase the max_discard_size, just afaik not documented anywhere, so not widely used. And even than it is not easy to reproduce...
I would leave it as it is for the sake of simplicity. :)

Thanks,
Jakub

Comment 6 Marek Suchanek 2018-08-09 10:31:54 EDT

OK, I'm adding "in rare cases" :-)

Comment 8 errata-xmlrpc 2018-08-16 10:19:04 EDT

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2018:2450

Note You need to log in before you can comment on or make changes to this bug.