Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1361182

Summary:	vdsm does not find nfs4 mount point with double slashes
Product:	[oVirt] vdsm	Reporter:	wodel <wodel.youchi>
Component:	General	Assignee:	Idan Shaby <ishaby>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Kevin Alon Goldblatt <kgoldbla>
Severity:	low	Docs Contact:
Priority:	unspecified
Version:	4.18.6	CC:	amureini, bugs, ishaby, nsoffer, tnisan, wodel.youchi
Target Milestone:	ovirt-4.0.4	Flags:	rule-engine: ovirt-4.0.z+ rule-engine: planning_ack+ rule-engine: devel_ack+ acanan: testing_ack+
Target Release:	4.18.12
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:	Previously, we didn't support NFS4 mounts that contained a double slash, like in the case of trying to mount an NFS4 directory that is exported with the fsid=0 option. Now it's possible to do that.	Story Points:	---
Clone Of:		Environment:
Last Closed:	2016-09-26 12:41:11 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	Storage	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description wodel 2016-07-28 12:59:38 UTC

Description of problem:

When using nfs4 with fsid=0 to change the root of the exported directories, the mounted share is displayed with double slashes, i.e
nfsserver # cat /etc/exports
/home/ovirt             *(fsid=0,rw,....)
/home/ovirt/engine_vm   *(rw,....)

hypervisor # mount -t nfs4 nfsserver.domain.tld:/engine_vm /mnt
hypervisor # df -h
nfsserver.domain.tld://engine_vm          656G    111G  513G  18% /mnt

even the mount command shows the double slashes.
hypervisor # cat /proc/mount
...
nfsserver.domain.tld://engine_vm on /rhev/data-center/mnt/nfsserver.domain.tld:_engine__vm type nfs4 (rw,relatime,vers=4.0,...)


When the hosted-engine --deploy is started, I get this error when using the above mount point (nfsserver.domain.tld:/engine_vm)
Failed to execute stage 'Environment customization': Connection to storage server failed

After some digging in vdsm logs, I found that vdsm searches for :
OSError: [Errno 2] Mount of `nfsserver.domain.tld:/engine_vm` at `/rhev/data-center/mnt/nfsserver.domain.tld:_engine__vm` does not exist

I think that vdsm can't find the mount point because of the double slashes.

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

OS : Fedora 23
oVirt : 4.0.1
vdsm-4.18.6-1.fc23.x86_64
Storage : nfs4 with fsid=0

How reproducible:
always

Steps to Reproduce:
1.configure nfs4 storage with fsid=0 to hide the root export.
2.install ovirt-hosted-engine-setup and start the deployment.
3.use the nfs4 export for the engine's storage

Actual results:
[ ERROR ] Failed to execute stage 'Environment customization': Connection to storage server failed

Expected results:
The mount is executed successfully.

Additional info:

Comment 1 Allon Mureinik 2016-08-08 13:00:35 UTC

This is pretty visible. Idan, please add some doctext.

Comment 2 Allon Mureinik 2016-08-10 08:21:35 UTC

Oops, the needinfo was intended for Idan, not Wodel. My apologies.

Comment 3 Nir Soffer 2016-08-10 15:09:29 UTC

This needs a backport to 4.0. I believe this is a regression introduced in 
https://gerrit.ovirt.org/55182.

Comment 4 Idan Shaby 2016-08-16 08:04:18 UTC

Please review the backports - patches 61580 and 61581.
Thanks

Comment 5 Kevin Alon Goldblatt 2016-09-11 13:29:15 UTC

Tested with the following code:
----------------------------------------
rhevm-4.0.4-0.1.el7ev.noarch
vdsm-4.18.12-1.el7ev.x86_64

Tested with the following scenario:

Steps to Reproduce:
1. Created an nfs mount as follows:

cat /etc/exports
/home/ovirt             *(fsid=0,rw,....)
/home/ovirt/engine_vm   *(rw,....)

hypervisor # mount -t nfs4 nfsserver.domain.tld:/engine_vm /mnt
hypervisor # df -h
nfsserver.domain.tld://engine_vm          656G    111G  513G  18% /mnt

even the mount command shows the double slashes.
hypervisor # cat /proc/mount
...
nfsserver.domain.tld://engine_vm on /rhev/data-center/mnt/nfsserver.domain.tld:_engine__vm type nfs4 (rw,relatime,vers=4.0,...)Actual results:
The first disk is now marked as bootable by default

2. Deployed the hosted engine installer and enterd the 

When the hosted-engine --deploy is started, I get NO error when using the above mount point (nfsserver.domain.tld:/engine_vm) and the storage domain is created successfully as part of a successful HE deployment


Expected results:



Moving to VERIFIED!