Bug 1397283

Summary:	[Disperse] : I/O errors on all FUSE mounts during rm -rf from multiple clients
Product:	[Community] GlusterFS	Reporter:	Ashish Pandey <aspandey>
Component:	disperse	Assignee:	Ashish Pandey <aspandey>
Status:	CLOSED DUPLICATE	QA Contact:
Severity:	high	Docs Contact:
Priority:	medium
Version:	mainline	CC:	aspandey, bturner, bugs, jahernan, nchilaka, rhinduja, rhs-bugs, storage-qa-internal
Target Milestone:	---
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:	1395161	Environment:
Last Closed:	2018-10-25 14:20:18 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:
Bug Depends On:	1395161, 1812789
Bug Blocks:

Comment 1 Ashish Pandey 2016-11-22 07:13:38 UTC

Steps -

1 - Create distributed disperse volume. 2 X (4+2)
2 - Kill 2 bricks from each subvolumes.
3 - mount this volume on 2 clients and create 1 directories from both the clients on mount point.
4 - Untar linux tar ball on these directories from different clients.
5 - On mount points of both the clients run "rm -rf *" .
6 - After some time you will see Input/Output Errors for few files/dir.

Comment 2 Worker Ant 2016-11-28 11:27:10 UTC

REVIEW: http://review.gluster.org/15946 (cluster/ec: Get/Set severity of error when inodelk fails) posted (#1) for review on master by Ashish Pandey (aspandey)

Comment 3 Ashish Pandey 2017-03-08 08:36:33 UTC

Previous patch needs to be reworked from scratch.

Comment 4 Xavi Hernandez 2018-10-25 14:20:18 UTC

The root cause of this problem is the same identified in bug #1578405, i.e. the deletion of entries while they are locked or being locked.

Closing this bug as a duplicate.

*** This bug has been marked as a duplicate of bug 1578405 ***