Bug 1477593

Summary: nfs-server fails to start when IPv6 disabled
Product: Red Hat Enterprise Linux 7 Reporter: Marko Myllynen <myllynen>
Component: nfs-utilsAssignee: Steve Dickson <steved>
Status: CLOSED DUPLICATE QA Contact: Yongcheng Yang <yoyang>
Severity: urgent Docs Contact:
Priority: urgent    
Version: 7.4CC: bruce.howells, dwysocha, eguan, jeharris, jiyin, joshua.megerman, jswensso, mvanderw, pdwyer, rmarigny, wbaudler, woodard, xzhou, yoyang, zlang
Target Milestone: rcKeywords: Regression
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: nfs-utils-1.3.0-0.50.el7 Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2017-08-07 06:07:56 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 Marko Myllynen 2017-08-02 12:31:50 UTC
Description of problem:
On RHEL 7.3 with nfs-utils-1.3.0-0.33.el7_3.x86_64 "systemctl start nfs-server" always worked reliably on systems which had IPv6 disabled with the kernel boot parameter ipv6.disable=1. Now on RHEL 7.4 with nfs-utils-1.3.0-0.48.el7.x86_64 on the same systems nfs-server fails almost always to start.

Here's an example run on an IPv6-less test system with nfs-utils-1.3.0-0.33.el7_3.x86_64:

# systemctl start nfs-server
# systemctl stop nfs-server 
# systemctl start nfs-server
# systemctl stop nfs-server 
# systemctl start nfs-server
# systemctl stop nfs-server 

On the same system doing merely "yum update nfs-utils" we then see with nfs-utils-1.3.0-0.48.el7.x86_64:

# systemctl start nfs-server
Job for nfs-server.service failed because the control process exited with error code. See "systemctl status nfs-server.service" and "journalctl -xe" for details.
zsh: exit 1     systemctl start nfs-server
# systemctl stop nfs-server 
# systemctl start nfs-server
# systemctl stop nfs-server 
# systemctl start nfs-server
Job for nfs-server.service failed because the control process exited with error code. See "systemctl status nfs-server.service" and "journalctl -xe" for details.
zsh: exit 1     systemctl start nfs-server
# systemctl stop nfs-server 

This is what is logged to /var/log/messages after a failure:

Aug  2 15:31:06 mars systemd: Starting Preprocess NFS configuration...
Aug  2 15:31:06 mars systemd: Started Preprocess NFS configuration.
Aug  2 15:31:06 mars systemd: Starting NFS Mount Daemon...
Aug  2 15:31:06 mars systemd: Starting NFSv4 ID-name mapping service...
Aug  2 15:31:06 mars systemd: Started NFSv4 ID-name mapping service.
Aug  2 15:31:06 mars rpc.mountd[8031]: Version 1.3.0 starting
Aug  2 15:31:06 mars systemd: Started NFS Mount Daemon.
Aug  2 15:31:06 mars systemd: Starting NFS server and services...
Aug  2 15:31:07 mars rpc.nfsd: rpc.nfsd: address family AF_INET6 not supported by protocol TCP
Aug  2 15:31:07 mars rpc.nfsd: rpc.nfsd: unable to set any sockets for nfsd
Aug  2 15:31:07 mars systemd: nfs-server.service: main process exited, code=exited, status=1/FAILURE
Aug  2 15:31:07 mars systemd: Failed to start NFS server and services.
Aug  2 15:31:07 mars systemd: Unit nfs-server.service entered failed state.
Aug  2 15:31:07 mars systemd: nfs-server.service failed.
Aug  2 15:31:07 mars systemd: Stopping NFS Mount Daemon...
Aug  2 15:31:07 mars systemd: Stopping NFSv4 ID-name mapping service...
Aug  2 15:31:07 mars rpc.mountd[8031]: Caught signal 15, un-registering and exiting.
Aug  2 15:31:07 mars systemd: Starting Notify NFS peers of a restart...
Aug  2 15:31:07 mars systemd: Stopped NFSv4 ID-name mapping service.
Aug  2 15:31:07 mars systemd: Stopped NFS Mount Daemon.
Aug  2 15:31:07 mars sm-notify[8046]: Version 1.3.0 starting
Aug  2 15:31:07 mars sm-notify[8046]: Already notifying clients; Exiting!
Aug  2 15:31:07 mars systemd: Started Notify NFS peers of a restart.

Going back to 1.3.0-0.33.el7_3.x86_64 with "yum downgrade nfs-utils" fixes it again.

Version-Release number of selected component (if applicable):
nfs-utils-1.3.0-0.48.el7.x86_64

Comment 7 Yongcheng Yang 2017-08-07 06:07:56 UTC
Seems we already have Bug 1450528 to track this issue.

Closing this one and please correct me if anyone concern.

*** This bug has been marked as a duplicate of bug 1450528 ***

Comment 8 Matt 2017-08-07 13:09:50 UTC
Anyway that BZ#1450528 can be made public (like this one is) so we can follow it's progress? We're having the same issues on our RHEL 7 systems and I was previously following this bugzilla entry so I could see what progress there is on a fix for this issue. We have an active subscription to RHEL.

Thanks.

Comment 9 Yongcheng Yang 2017-08-08 03:17:06 UTC
(In reply to Matt from comment #8)
Hello Matt, sorry for not considering this issue before.

BZ#1450528 was filed by a customer and there already have some
investigation there, so I duplicate this one to it. However, we may
need their approve before making it public.

Anyway, I'll continue to update progress here for your information.
Thanks for understanding.

Comment 10 Joshua Megerman 2017-08-14 16:57:21 UTC
I'm having the same issue.  I can start the nfs-server service manually, but it fails to start properly on server boot.  This is a significant issue because we have servers that get rebooted automatically overnight with the expectation that no manual intervention will be needed to bring everything back up.

I guess I'll downgrade my nfs-utils for the time being and versionlock them until I hear more.  Would it be better if I opened a case with my RHEL subscription to get notifications that way?

Thanks.

Comment 13 Wolfgang Baudler 2017-09-25 12:45:41 UTC
Same problem here. Any progress on fixing this or providing a workaround? Thanks.

Comment 14 Yongcheng Yang 2017-10-09 09:03:23 UTC
(In reply to Yongcheng Yang from comment #9)
> (In reply to Matt from comment #8)
> Hello Matt, sorry for not considering this issue before.
> 
> BZ#1450528 was filed by a customer and there already have some
> investigation there, so I duplicate this one to it. However, we may
> need their approve before making it public.
> 
> Anyway, I'll continue to update progress here for your information.
> Thanks for understanding.

Just for your information, bug 1450528 has been fixed in version
nfs-utils-1.3.0-0.50.el7 now.

Also it has been copied as 7.4 z-stream (EUS) bug #1498959.

Comment 15 Matt 2017-10-10 13:57:19 UTC
Thanks for the update!!

Any idea on an ETA for when this nfs-utils package update will be made available for the normal RHEL repo/subscription?

Comment 16 Yongcheng Yang 2017-10-20 06:39:51 UTC
The zstream package (nfs-utils-1.3.0-0.48.el7_4) which has fixed
this issue should be available now (2017-Oct-19 shipped).

Comment 17 Matt 2017-10-20 14:48:10 UTC
Yes, thank you! I noticed it came out yesterday and can confirm that the update does indeed fix the issues we've been seeing!

Thanks!

Comment 18 Wolfgang Baudler 2017-10-20 14:58:46 UTC
Same here.