autofs5 fails to install: post-installation script returned error exit status 1

[Dmitri Fedorov]

$ sudo apt-get install autofs5
...
Setting up autofs5 (5.0.6-0ubuntu4) ...
start: Job failed to start
invoke-rc.d: initscript autofs, action "start" failed.
dpkg: error processing autofs5 (--configure):
 subprocess installed
Errors were encountered while processing:
 autofs5
E: Sub-process /usr/bin/dpkg returned an error code (1)

and here is the apport bug file:

ProblemType: Bug
ApportVersion: 1.94.1-0ubuntu2
Architecture: amd64
Date: Sun Mar 11 16:42:11 2012
Dependencies:
 adduser 3.113ubuntu2
 base-passwd 3.5.24
 busybox-initramfs 1:1.18.5-1ubuntu2
 coreutils 8.13-3ubuntu2
 cpio 2.11-7ubuntu1
 debconf 1.5.41ubuntu2
 debianutils 4.2.1ubuntu1
 dpkg 1.16.1.2ubuntu5
 findutils 4.4.2-4ubuntu1
 gcc-4.6-base 4.6.3-1ubuntu3
 ifupdown 0.7~beta2ubuntu6
 initramfs-tools 0.99ubuntu11
 initramfs-tools-bin 0.99ubuntu11
 initscripts 2.88dsf-13.10ubuntu10
 insserv 1.14.0-2.1ubuntu1
 iproute 20111117-1ubuntu1
 klibc-utils 1.5.25-1ubuntu1
 libacl1 2.2.51-5ubuntu1
 libattr1 1:2.4.46-5ubuntu1
 libblkid1 2.20.1-1ubuntu2
 libbz2-1.0 1.0.6-1
 libc-bin 2.15-0ubuntu5
 libc6 2.15-0ubuntu5
 libdb5.1 5.1.25-11build1
 libdbus-1-3 1.4.18-1ubuntu1
 libdrm-intel1 2.4.30-1ubuntu1
 libdrm-nouveau1a 2.4.30-1ubuntu1
 libdrm-radeon1 2.4.30-1ubuntu1
 libdrm2 2.4.30-1ubuntu1
 libffi6 3.0.11~rc1-5
 libgcc1 1:4.6.3-1ubuntu3
 libglib2.0-0 2.31.20-0ubuntu2
 libklibc 1.5.25-1ubuntu1
 liblzma5 5.1.1alpha+20110809-3
 libmount1 2.20.1-1ubuntu2
 libncurses5 5.9-4
 libncursesw5 5.9-4
 libnih-dbus1 1.0.3-4ubuntu9
 libnih1 1.0.3-4ubuntu9
 libpam-modules 1.1.3-7ubuntu2
 libpam-modules-bin 1.1.3-7ubuntu2
 libpam0g 1.1.3-7ubuntu2
 libpciaccess0 0.12.902-1
 libpcre3 8.12-4
 libplymouth2 0.8.2-2ubuntu28
 libpng12-0 1.2.46-3ubuntu2
 libselinux1 2.1.0-4.1ubuntu1
 libslang2 2.2.4-3ubuntu1
 libtinfo5 5.9-4
 libudev0 175-0ubuntu6
 libuuid1 2.20.1-1ubuntu2
 libxml2 2.7.8.dfsg-5.1ubuntu4
 lsb-base 4.0-0ubuntu20
 makedev 2.3.1-89ubuntu2
 module-init-tools 3.16-1ubuntu2
 mount 2.20.1-1ubuntu2
 mountall 2.35
 multiarch-support 2.15-0ubuntu5
 ncurses-bin 5.9-4
 passwd 1:4.1.4.2+svn3283-3ubuntu4
 perl-base 5.14.2-6ubuntu1
 plymouth 0.8.2-2ubuntu28
 procps 1:3.2.8-11ubuntu6
 sed 4.2.1-9
 sensible-utils 0.0.6ubuntu2
 sysv-rc 2.88dsf-13.10ubuntu10
 sysvinit-utils 2.88dsf-13.10ubuntu10
 tar 1.26-4
 tzdata 2012b-1
 ucf 3.0025+nmu2ubuntu1
 udev 175-0ubuntu6
 upstart 1.4-0ubuntu8
 util-linux 2.20.1-1ubuntu2
 xz-utils 5.1.1alpha+20110809-3
 zlib1g 1:1.2.3.4.dfsg-3ubuntu4
DistroRelease: Ubuntu 12.04
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Beta amd64 (20120227)
Package: autofs5 5.0.6-0ubuntu4
PackageArchitecture: amd64
ProcEnviron:
 LC_PAPER=en_DK.UTF-8
 LC_ADDRESS=en_DK.UTF-8
 LC_MONETARY=en_DK.UTF-8
 TERM=screen
 SHELL=/bin/bash
 LC_NUMERIC=en_DK.UTF-8
 LC_TELEPHONE=en_DK.UTF-8
 LANG=en_US.UTF-8
 LC_MEASUREMENT=en_DK.UTF-8
 LANGUAGE=en_US:en
 LC_TIME=en_DK.UTF-8
 LC_NAME=en_DK.UTF-8
ProcVersionSignature: Ubuntu 3.2.0-18.29-generic 3.2.9
SourcePackage: autofs5
Tags: precise
Uname: Linux 3.2.0-18-generic x86_64
UpgradeStatus: No upgrade log present (probably fresh install)

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Clint Byrum (clint-fewbar)]

Hi Dmitri. Can you please include the output of

grep autofs /var/log/syslog

and if it exists, /var/log/upstart/autofs.log ?

Thanks!

[Dmitri Fedorov (dmon137)]

@Clint Byrum:

$ grep autofs /var/log/syslog | tail
Mar 11 16:27:44 molveno kernel: [64449.010472] init: wait-for-state (autofs5ypbind) main process (19697) terminated with status 100
Mar 11 16:27:44 molveno kernel: [64449.020729] init: autofs pre-start process (19694) terminated with status 1
Mar 11 16:56:55 molveno kernel: [66196.495099] init: wait-for-state (autofs5ypbind) main process (21150) terminated with status 100
Mar 11 16:56:55 molveno kernel: [66196.500478] init: autofs pre-start process (21147) terminated with status 1
Mar 11 17:44:16 molveno kernel: [69031.679746] init: wait-for-state (autofs5ypbind) main process (22101) terminated with status 100
Mar 11 17:44:16 molveno kernel: [69031.686177] init: autofs pre-start process (22098) terminated with status 1
Mar 11 17:45:26 molveno kernel: [69101.285300] init: wait-for-state (autofs5ypbind) main process (22563) terminated with status 100
Mar 11 17:45:26 molveno kernel: [69101.290567] init: autofs pre-start process (22560) terminated with status 1
Mar 11 17:47:06 molveno kernel: [69201.474770] init: wait-for-state (autofs5ypbind) main process (23392) terminated with status 100
Mar 11 17:47:06 molveno kernel: [69201.481256] init: autofs pre-start process (23389) terminated with status 1

$ sudo cat /var/log/upstart/autofs.log | tail
start: Job failed to start
start: Job failed to start
start: Job failed to start
start: Job failed to start
start: Job failed to start
start: Job failed to start
start: Job failed to start
start: Job failed to start
start: Job failed to start
start: Job failed to start

[Dmitri Fedorov (dmon137)]

I have just realised that despite all the error messages, the autofs system works and mounts the NFS filesystem just fine:

$ ps -A | grep auto
 9804 ? 00:00:00 automount

$ mount | grep nfs
lirs10:/a/home/lirs10b/fedorov on /usr/users/fedorov type nfs (rw,addr=10.12.1.78)

nevertheless:

$ sudo dpkg --configure -a
Setting up autofs5 (5.0.6-0ubuntu4) ...
start: Job failed to start
invoke-rc.d: initscript autofs, action "start" failed.
dpkg: error processing autofs5 (--configure):
 subprocess installed post-installation script returned error exit status 1
Errors were encountered while processing:
 autofs5

Have I messed up my system somehow?

[Dmitri Fedorov (dmon137)]

Ok, I can reproduce the bug any time now:

$ sudo apt-get install autofs5
...
Setting up autofs5 (5.0.6-0ubuntu4) ...
start: Job failed to start
invoke-rc.d: initscript autofs, action "start" failed.
dpkg: error processing autofs5 (--configure):
 subprocess installed post-installation script returned error exit status 1
Errors were encountered while processing:
 autofs5
E: Sub-process /usr/bin/dpkg returned an error code (1)

What happens---It seems---is that autofs fails to start via start utility:

$ sudo start autofs
start: Job failed to start
$ cd ~fedorov
bash: cd: /usr/users/fedorov: No such file or directory

However, if I manually start automount, everything seems to work just fine:

$ sudo automount
$ cd ~fedorov
/usr/users/fedorov$

So the bug seems to be not in autofs5 but rather in Upstart, no?

Should I mention also that the root filesystem on this box is on a raid0?

[Clint Byrum (clint-fewbar)]

Dmitri, thanks for the response. Can you also do

dpkg -l nis

And show the output of

status ypbind

Also if its 'stop/waiting'

then

start ypbind

and then show the output of 'cat /var/log/upstart/ypbind.log'

Thanks!!

[Dmitri Fedorov (dmon137)]

@Clint:

$ dpkg -l nis
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name Version Description
+++-==============-==============-============================================
ii nis 3.17-32ubuntu3 clients and daemons for the Network Informat

$ status ypbind
ypbind start/running, process 857

$ sudo cat /var/log/upstart/ypbind.log
$

$ sudo zcat /var/log/upstart/ypbind.log.*.gz | tail
Binding to YP server .done
Setting NIS domainname to: ifayp
Binding to YP server .....backgrounded
Setting NIS domainname to: ifayp
Binding to YP server ....done
Setting NIS domainname to: ifayp
Binding to YP server .....backgrounded
Setting NIS domainname to: ifayp
Binding to YP server .....backgrounded
Binding to YP server .done

[Clint Byrum (clint-fewbar)]

Thanks again. Thats very odd, as the autofs job should exit immediately when wait-for-state sees that ypbind is already running. I will spin up a precise VM here and see if I can reproduce.

[Clint Byrum (clint-fewbar)]

I *think* this is a duplicate of bug #958704 , which was fixed in version 5.0.6-0ubuntu5

Can any of the affected users confirm that the final released precise version does not have the reported problem?