Bug#622394: Fix works on NFS clients but breaks NFS servers

To: 622394@bugs.debian.org
Subject: Bug#622394: Fix works on NFS clients but breaks NFS servers
From: Chris Butler <cbutler@zedcore.com>
Date: Wed, 29 Oct 2014 11:06:18 +0000 (GMT)
Message-id: <[🔎] 604020462.280709.1414580778654.JavaMail.zimbra@zedcore.com>
Reply-to: Chris Butler <cbutler@zedcore.com>, 622394@bugs.debian.org
In-reply-to: <1420585642.279859.1414580446715.JavaMail.zimbra@zedcore.com>

Hi,

I applied the fix given in message #10 to a few of our Jessie systems. I can confirm that it fixes the boot order problem on NFS clients that mount remote filesystems on boot (before the fix, it was trying to mount an NFS /home directory before rpcbind was started, which failed).

However, the same fix seems to break NFS servers because it tries to start "statd" before "rpcbind":

Oct 29 09:48:49 drbl nfs-common[140]: Starting NFS common utilities: statd
Oct 29 09:48:49 drbl nfs-common[140]: Not starting: portmapper is not running ... (warning).

This failure to start later manifested as "lockd: cannot monitor <client>" errors.

-- 
Chris Butler
Zedcore Systems Ltd

Reply to:

Prev by Date: Bug#767219: Restarting NFS server while clients are connected causes locking errors
Next by Date: Bug#766802: linux-image-amd64: rt2800usb driver does not include 1b75:a200 USB id's as a supported device
Previous by thread: Bug#767219: Restarting NFS server while clients are connected causes locking errors
Next by thread: Bug#767239: [linux] Regression: T420 docking station headphone plug not working automatically
Index(es):
- Date
- Thread