This appears to be the standard litany of the new user, but I have been
trying everything I can think of to get the amandad to start on my initial
backup machine running RedHat 7.1. I have gone through the archives and
tried all of the steps suggested in all of the archived articles on this
subject without success. The one extenuating circumstance that I have is
that I have run the Bastille program on this machine before I started to
configure Amanda. To the best of my knowledge all access is opened up
though. As it sits I have tried/checked the following:
1: I can run the program using the following: su - amanda -c
"/usr/lib/amanda/amandad"
This results in the following output in the debug file:
amandad: debug 1 pid 22208 ruid 33 euid 33 start time Fri Aug 24 20:30:10
2001
/usr/lib/amanda/amandad: version 2.4.2p2
/usr/lib/amanda/amandad: build: VERSION="Amanda-2.4.2p2"
/usr/lib/amanda/amandad: BUILT_DATE="Thu Apr 5 09:18:00 EDT 2001"
/usr/lib/amanda/amandad: BUILT_MACH="Linux porky.devel.redhat.com
2.2.17-8smp #1 SMP Fri Nov 17 16:12:17 EST 2000 i686 unknown"
/usr/lib/amanda/amandad: CC="gcc"
/usr/lib/amanda/amandad: paths: bindir="/usr/bin" sbindir="/usr/sbin"
/usr/lib/amanda/amandad: libexecdir="/usr/lib/amanda"
mandir="/usr/share/man"
/usr/lib/amanda/amandad: AMANDA_TMPDIR="/tmp/amanda"
AMANDA_DBGDIR="/tmp/amanda"
/usr/lib/amanda/amandad: CONFIG_DIR="/etc/"
/usr/libamanda/amandad: DEFAULT_TAPE_SERVER="localhost"
/usr/lib/amanda/amanda" DEV_PREFIX="/dev/"
/usr/lib/amanda/amandad: RDEV_PREFIX="/dev/" DUMP="/sbin/dump"
/usr/lib/amanda/amandad: RESTORE="/sbin/restore"
SAMBA_CLIENT="/usr/bin/smbclient"
/usr/lib/amanda/amandad: GNUTAR="/bin/tar"
COMPRESS_PATH="/usr/bin/gzip"
/usr/lib/amanda/amandad: UNCOMPRESS_PATH="/usr/bin/gzip"
MAILER="/usr/bin/Mail"
/usr/lib/amanda/amandad:
listed_incr_dir="/var/lib/amanda/gnutar-lists"
/usr/lib/amanda/amandad: defs: DEFAULT_SERVER="localhost"
DEFAULT_CONFIG="DailySet1"
/usr/lib/amanda/amandad: DEFAULT_TAPE_DEVICE="/dev/null" HAVE_MMAP
HAVE_SYSVSHM
/usr/lib/amanda/amandad: LOCKING=POSIX_FCNTL SETPGRP_VOID DEBUG_CODE
/usr/lib/amanda/amandad: AMANDA_DEBUG_DAYS=4 BSD_SECURITY
USE_AMANDAHOSTS
/usr/lib/amanda/amandad: CLIENT_LOGIN="amanda" FORCE_USERID HAVE_GZIP
/usr/lib/amanda/amandad: COMPRESS_SUFFIX=".gz"
COMPRESS_FAST_OPT="--fast"
/usr/lib/amanda/amandad: COMPRESS_BEST_OPT="--best"
UNCOMPRESS_OPT="-dc"
amandad: dgram_recv: timeout after 30 seconds
amandad: error receiving message: timeout error receiving message: timeout
amandad: pid 22208 finish time Fri Aug 24 20:30:40 2001
2: I have checked to see that the TCPChains are setup to allow me to run
this as network request. The current value in the hosts.allow file is:
amandad : 127.0.0.1 207.173.31.0/255.255.250.0 : spawn
(/usr/sbin/safe_finger -l @%h | /bin/mail -s "Amandad attempt noted %d-%h"
root) & :ALLOW
I have tried it both with and without the Spawn command. It was only added
to verify that I was getting the request in.
3. I have tried replacing the amandad with a script as was suggested,
verifying by hand that it would run before placing it in the
/etc/xinetd.d/amanda file.
4: The /etc/xinetd.d/amanda file contains:
# default: off
# description: The client for the Amanda backup system.\
# This must be on for systems being backed up\
# by Amanda.
service amanda
{
socket_type = dgram
protocol = udp
wait = yes
user = amanda
group = disk
server = /usr/lib/amanda/amandad
# server = /bin/truss amandad -fo /tmp/amandad.truss
/usr/lib/amanda/amandad
# server =/tmp/amandadtest
disable = no
}
5. I have tried both the truss listed in the amanda xinetd.d file above and
the amandatest script as described by someone earlier in the list. I had to
convert it to bash as I don't have ksh loaded on this box. The contents of
that script are:
#!/bin/bash
exec > /tmp/amandad.out.$$ 2>&1
echo "$(date): starting amandad"
/usr/lib/amanda/amandad
echo "$(date): amandad done: status is $?"
exit 0
This command runs by hand correctly as well and leaves a status file behind.
Nothing that I do results in amandad being run or any diagnostic information
being left behind in the /tmp (when using the test script as suggested) nor
/tmp/amanda when trying to invoke it straight, other than that the amandad
service was looping and xinetd was killing it. The spawn command does send
email and tell me that an attempt was made though.
I saw a hint that under Solaris it was necessary to add the paths to the
amanda shared libraries to the startup scripts, but I am using the supplied
2.4.2Ps version that came with the 7.1 release, and the libraries are in
/lib and /usr/lib, which are in the standard path.
Need new suggestions. Does anybody else have this running under RedHat 7.1
on a machine that they ran Bastille on?
===========================================================
Mark A. Holm - Director of Network Engineering & Operations
Inherent.com, Inc.
2140 SW Jefferson St. Phone: (503) 224-6751
Suite 200 extension 236
Portland, OR 97201 Fax: (503) 224-8872
<http://www.inherent.com> <mailto:[EMAIL PROTECTED]>
