Bug 18441 – Major problems in /etc/init.d/afs script

• Product: Gentoo Linux
• Status: RESOLVED FIXED

[Description Chris Snyder 2003-03-29 14:30:47 UTC]

I have three machines that have problems starting AFS using the /etc/init.d/afs startup 
script. They all return an error like follows: 
 
 * Starting AFS services... 
Failed to load AFS client, not starting AFS services. 
/sbin/runscript.sh: [: Error Starting AFS client: integer expression expected                                         
[ !! ] 
 
/sbin/runscript.sh: return: bad non-numeric arg `Error' 
afsd: All AFS daemons started. 
afsd: Can't mount AFS on /afs(2) 
 * Error starting AFS                                                                                                 [ !! ] 
 
While this looks similar to another bug reported earlier, on closer inspection it appears 
to be a different error. 
 
When I load the AFS kernel module and start afsd separately, everything works fine, 
so I know that my configuration is fine. 
 
Also, another problem with the startup script is that it always ensures that the 
/usr/vice/cache partition is ext2. However, if a memory cache is being used, this 
partition will not exist, causing the script to fail. My solution has been to comment out 
the function in the startup script that checks for this, but there should be a more 
elegant solution. 

Reproducible: Always
Steps to Reproduce:
1. Try to start AFS using the /etc/init.d/afs script 
Actual Results:  
Error 

Expected Results:  
Started AFS 

The systems are all Gentoo 1.2 systems. I have not had the problem occur on a 1.4 
system.

[Comment 1 Brad Cowan (RETIRED) 2003-05-05 01:49:30 UTC]

is it possible your 1.2 machines need baselayout updated, or the config files that changed after emerging baselayout were not updated?

[Comment 2 Chris Snyder 2003-05-05 01:55:49 UTC]

Doubtful. emerge -u world was done fairly recently on all the machines. I'm also not 
exactly sure when this started happening - they were working fine, but then they 
rebooted due to a power failure (after ~250 days uptime) and AFS didn't start again. I 
think I may have upgraded AFS a couple of times in between it being restarted. I'll 
check the config files for any errors, though I'll have to be careful, as these are 
production systems serving http directly off of AFS. 

[Comment 3 Brad Cowan (RETIRED) 2003-05-05 02:00:57 UTC]

did u etc-update after emerge -u world to replace the updated config files? it's just weird that it works on a 1.4 box and not 1.2 because the main difference is really the compiler 

[Comment 4 Brad Cowan (RETIRED) 2003-05-05 13:11:47 UTC]

I'm assigning to our AFS expert if thats alright, rphillips.

[Comment 5 Chris Snyder 2003-05-13 18:59:16 UTC]

All configuration files are up to date. I'm not sure if the fact that they're 1.2 boxen is really the 
problem, but that's the pattern I'm noticing. I'm attaching two emerge infos: one from a system 
that's not starting afs properly, and another from a system where afs is working properly. 
 
Here's a system where it's not working: 
Portage 2.0.47-r10 (default-1.0, gcc-2.95.3, glibc-2.2.5-r4,2.2.5-r8) 
================================================================= 
System uname: 2.4.20 i686 AMD Duron(tm) processor 
GENTOO_MIRRORS="ftp://mirror.iawnet.sandia.gov/pub/gentoo" 
CONFIG_PROTECT="/etc /var/qmail/control /usr/share/config /usr/kde/2/share/config 
/usr/kde/3/share/config" 
CONFIG_PROTECT_MASK="/etc/gconf /etc/afs/C /etc/afs/afsws /etc/env.d" 
PORTDIR="/usr/portage" 
DISTDIR="/usr/portage/distfiles" 
PKGDIR="/usr/portage/packages" 
PORTAGE_TMPDIR="/var/tmp" 
PORTDIR_OVERLAY="/usr/local/portage" 
USE="x86 libg++ libwww gdbm slang mmx 3dnow sse -3dfx -voodoo3 -matrox -apm -pcmcia 
-pnp -pda -gphoto2 cups -X -dga -opengl -xv -fbcon -directfb -kde -qt -qtmt -arts -gnome -gtk 
-bonobo -evo -gb -gtkhtml motif tcltk aalib imlib ncurses readline -sdl -svga lcms gif jpeg png tiff 
avi mpeg quicktime -alsa -esd -mikmod -oss -nas -ggi -dvd -xmms -oggvorbis -flash encode 
-ipv6 -snmp -ggz pam ssl crypt -socks5 imap -ldap tcpd -xface spell -truetype xml2 pdflib 
plotutils -tetex -guile java perl python berkdb mysql innodb -acl afs samba -nls doc -gpm" 
COMPILER="" 
CHOST="i586-pc-linux-gnu" 
CFLAGS="-mcpu=k6 -O3 -pipe" 
CXXFLAGS="-mcpu=k6 -O3 -pipe" 
ACCEPT_KEYWORDS="x86" 
MAKEOPTS="-j2" 
AUTOCLEAN="yes" 
SYNC="rsync://server2.mvpsoft.servers/gentoo-portage" 
FEATURES="sandbox ccache" 
 
Here's a system where it is working: 
Portage 2.0.47-r10 (default-x86-1.4, gcc-3.2.2, glibc-2.3.1-r4) 
================================================================= 
System uname: 2.4.20 i686 AMD Athlon(tm) XP 1600+ 
GENTOO_MIRRORS="http://www.ibiblio.org/pub/Linux/distributions/gentoo" 
CONFIG_PROTECT="/etc /var/qmail/control /usr/kde/2/share/config /usr/kde/3/share/config 
/usr/X11R6/lib/X11/xkb /usr/kde/3.1/share/config /usr/share/config" 
CONFIG_PROTECT_MASK="/etc/afs/C /etc/afs/afsws /etc/gconf /etc/env.d" 
PORTDIR="/usr/portage" 
DISTDIR="/afs/.mvpsoft.com/common/distfiles" 
PKGDIR="/usr/portage/packages" 
PORTAGE_TMPDIR="/var/tmp" 
PORTDIR_OVERLAY="/usr/local/portage" 
USE="x86 oss 3dnow apm gnome mikmod mmx nls zlib gdbm berkdb tcpd esd gtk cdr cups 
slp X dga opengl xv fbcon kde qtmt arts qt motif tcltk aalib imlib ncurses readline sdl svga lcms 
gif jpeg png tiff avi mpeg quicktime alsa dvd xmms oggvorbis flash encode pam ssl crypt imap 
mozilla mozirc mozxmlterm spell truetype xml2 pdflib tetex guile java perl python ruby slang 
libwww libg++ pic mysql acl gpm" 
COMPILER="gcc3" 
CHOST="i686-pc-linux-gnu" 
CFLAGS="-O2 -mcpu=i686 -pipe" 
CXXFLAGS="-O2 -mcpu=i686 -pipe" 
ACCEPT_KEYWORDS="x86" 
MAKEOPTS="-j2" 
AUTOCLEAN="yes" 
SYNC="rsync://server2.mvpsoft.servers/gentoo-portage" 
FEATURES="sandbox ccache" 
 
I guess I'm going to have to add afs startup routines to rc.local for now... 

[Comment 6 Andrew Cooks (RETIRED) 2003-12-08 07:28:59 UTC]

I suppose this problem has gone away by now?

Can it be resolved as WONTFIX if it's still broken?

[Comment 7 Ryan Phillips (RETIRED) 2003-12-08 08:32:26 UTC]

I've added the patch from #26213.  I believe this will solve the problem.