Bug 998024

Summary: nfsd sometimes grants delegations too soon following conflicting open requests
Product: Red Hat Enterprise Linux 6 Reporter: Scott Mayhew <smayhew>
Component: kernelAssignee: J. Bruce Fields <bfields>
Status: CLOSED ERRATA QA Contact: JianHong Yin <jiyin>
Severity: medium Docs Contact:
Priority: medium    
Version: 6.4CC: bfields, ccui, cww, eguan, moshiro, smayhew, spanjikk, steved
Target Milestone: rc   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: kernel-2.6.32-489.el6 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
: 1111227 (view as bug list) Environment:
Last Closed: 2014-10-14 05:22:32 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:    
Bug Blocks: 994246, 1056239, 1111227    

Description Scott Mayhew 2013-08-16 18:48:37 UTC
Description of problem:
nfsd sometimes grants delegations too soon following conflicting open requests

Version-Release number of selected component (if applicable):
N/A

How reproducible:
Easily reproducible.

Steps to Reproduce:
Client 1:

1. open a file and leave it open (ensures we already have some state
established on the server when it's deciding whether to grant us a
delegation)
# vi /mnt/file1

2. switch to another terminal window and run the following
# touch /mnt/file2
# while :; do cat /mnt/file2; done
Client 1 gets a read delegation on in the OPEN reply for file2 in the
above command.

Client 2:
# date >> /mnt/file2

Actual results:
The command on client 2 never finishes.  What the server returns
NFS4ERR_DELAY to client 2 and recalls the delegation from client 1.
Client 1 returns the delegation to the server and immediately sends a
new OPEN, for which the server grants it another read delegation.  This
happens before client 2 "retries" the OPEN, so it's just going to wind
up getting NFS4ERR_DELAY again.

Expected results:
Client 2 should be able to open the file for writing after returning from its delay before Client 1 is granted another read delegation.

Additional info:
nfsd does currently track track the fact that there was a conflicting open via the nfs4_file->fi_had_conflict field... the problem is that the nfs4_file goes away when the reference count goes to zero.

Comment 10 J. Bruce Fields 2014-06-18 21:29:18 UTC
Note Neil Brown fixed this with 6282cd565553 "NFSD: Don't hand out delegations for 30 seconds after recalling them.", soon to be upstream.

Comment 13 Rafael Aquini 2014-07-08 02:19:19 UTC
Patch(es) available on kernel-2.6.32-489.el6

Comment 22 errata-xmlrpc 2014-10-14 05:22:32 UTC
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.

http://rhn.redhat.com/errata/RHSA-2014-1392.html