SOLVED NFS unavailable on one interface

[Chevalier]

Hello there. I hope you can eventually help me solve my problem.

I have a multihomed Freenas Server, generic hardware (FreeNAS-9.10.1-U2 (f045a8b)) with two links aggregates connected to different VLANS.

I want to share a dataset with NFS for ESX servers to make some ISOs available, on one of the interfaces.

The network setting is as follow :

Code:

lagg0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
   options=9b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM>
   ether 00:30:48:d5:17:d1
   inet 10.117.72.14 netmask 0xfffffc00 broadcast 10.117.75.255
   nd6 options=9<PERFORMNUD,IFDISABLED>
   media: Ethernet autoselect
   status: active
   laggproto lacp lagghash l2,l3,l4
   laggport: em1 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>
   laggport: em2 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>

lagg1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
   options=9b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM>
   ether 00:30:48:d5:17:d0
   inet 192.168.0.11 netmask 0xffffff00 broadcast 192.168.0.255
   nd6 options=9<PERFORMNUD,IFDISABLED>
   media: Ethernet autoselect
   status: active
   laggproto lacp lagghash l2,l3,l4
   laggport: em0 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>
   laggport: em3 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>

The ESX servers are on the 192.168.0.0/24 network. From them, I have no problem reaching the Freenas server, I can ping and ssh, no problem.

Problem : NFS is not reachable on the lagg1 interface. If I try to mount the share, it fails everytime.

This setup worked with the 9.2 and 9.3 versions !

A showmount, from another client on the same network shows :

Code:

# showmount -e 192.168.0.11
showmount: 192.168.0.11: RPC: Rpcbind failure - RPC: Unable to receive

If I check with nmap :

Code:

# nmap 192.168.0.11
Host is up (0.00033s latency).
Not shown: 995 closed ports
PORT  STATE SERVICE
22/tcp  open  ssh
80/tcp  open  http
139/tcp  open  netbios-ssn
445/tcp  open  microsoft-ds
2049/tcp open  nfs

Port 111 (rpcbind) is not open.

However, if I check the other address :

Code:

# nmap 10.117.72.14
Host is up (0.00033s latency).
Not shown: 994 closed ports
PORT  STATE SERVICE
22/tcp  open  ssh
80/tcp  open  http
111/tcp  open  rpcbind
139/tcp  open  netbios-ssn
445/tcp  open  microsoft-ds
2049/tcp open  nfs

A wild rpcbind appears...

And on this address, NFS works with no problem.

check on the Freenas server :

Code:

# netstat -na | grep 111
tcp4  0  0 10.117.72.14.111  *.*  LISTEN
tcp4  0  0 127.0.0.1.111  *.*  LISTEN
tcp6  0  0 ::1.111  *.*  LISTEN

It does not listen on lagg1 with tcp4.

If I look for the process :

Code:

# ps -aufx | grep rpcbind
root  93112  0.0  0.0  29160  2588  -  Is  11:07AM  0:00.01 /usr/sbin/rpcbind -h 10.117.72.14

I tried to check/uncheck both addresses in the gui, stoppping/restarting the service, to no avail, I can't make the "-h 10.117.72.14" go away.

I'm sure it worked with Freenas 9.2. I have another server, FreeNAS-9.2.1.9-RELEASE-x64 (2bbba09), where it works as expected : unchecking all the interfaces, rpcbind has no argument. Checking both, it gets two "-h" options.

How can I make it so that rpcbind gets no argument?

In a normal FreeBSD system, I'd just edit /etc/rc.conf, but here, I don't know.

I don't even know why it selects this interface.

 

[dlavigne]

Please create a bug report at bugs.freenas.org and post the issue number here.

 

[Chevalier]

I created the bug report : https://bugs.freenas.org/issues/23750

 

[Chevalier]

I solved it myself.

The problem was a left-over configuration of a rc.conf tunable in the gui of rpcbind_flags

Once the tunable removed, rpcbind worked as expected.