Hi all

I have Solaris 9 (latest patch cluster), Amanda 2.4.3 (and also tried 2.4.2)
and gcc 3.2.  Amanda compiles without problems (although it can't find
readline library, but I don't need it anyway) and installs without problems.
Amcheck runs without any problems (and so does amlabel).

amdump seems to start up okay and all the dumper and sendsize processes seem
to start, but then "driver" core dumps and the following is reported in the
amdump log file:

"HEAPS OF driver, taper and dumper messages here"
.....and then

driver: adding holding disk 0 dir /amholding2 size 7680000
reserving 7680000 out of 7680000 for degraded-mode dumps
driver: start time 7.379 inparallel 4 bandwidth 2400 diskspace 7680000 dir
OBSOLETE
datestamp 20021209 driver: drain-ends tapeq LFFO big-dumpers ttt
taper: read label `burnet1' date `20021209'
taper: wrote label `burnet1' date `20021209'
driver: result time 30.925 from taper: TAPER-OK
taper: DONE [idle wait: 0.159 secs]
Arithmetic Exception - core dumped
taper: writing end marker. [burnet1 OK kb 0 fm 0]
amdump: end at Mon Dec  9 21:33:55 EST 2002

The Arithmetic Exception error is I assume where something goes horribly
wrong.  A check on the core file in /tmp/amanda shows : core20021209:   ELF
32-bit MSB core file SPARC Version 1, from 'driver'

I have also checked a couple of previous versions of Amanda with the same
result.  It's really weird as everything closes down nicely so that I don't
even have to run amcleanup or anything either.  Here is the relevant parts
of the amanda.conf file:

org "burnet"
mailto "pauls AT loki.burnet.edu DOT au"
dumpcycle 0 days
runspercycle 1
tapecycle 1 tapes
#dumpuser
#printer
tapedev "/dev/rmt/0hn"
#rawtapedev
"/devices/iommu@0,10000000/sbus@0,10001000/espdma@4,8400000/esp@4,880000
0/st@4 ,0:"
#rawtapedev "/devices/pci@8,700000/scsi@3/st@4,0:n"
rawtapedev "/devices/pci@8,700000/scsi@3/st@4,0:hn"
#tpchanger
changerdev "/dev/null"
changerfile "/usr/local/etc/amanda/DailySet1/changer.conf"
runtapes 1
labelstr "^burnet[0-9][0-9]*$"
tapetype burnet-AIT
etimeout 300
netusage 1000 Kbps
inparallel 4
maxdumps 2
bumpsize 20 Mb
bumpmult 4
bumpdays 1
diskfile "/usr/local/etc/amanda/disklist"
infofile "/usr/local/etc/amanda/DailySet1/curinfo"
logdir "/usr/local/etc/amanda/DailySet1"
indexdir "/usr/local/etc/amanda/DailySet1/index"
tapelist "/usr/local/etc/amanda/tapelist"

holdingdisk hd2 {
   comment "main holding disk"
   directory "/amholding2"
   use 7500 Mb
   chunksize 1
}

define tapetype burnet-AIT {
    comment "just produced by tapetype program"
    length 32007 mbytes
    filemark 893 kbytes
    speed 1737 kps

All that gets produced in holdingdisk is the directory with the date and
thats all.
Sometimes a rundump.debug file gets produced with the following:
rundump: debug 1 pid 903 ruid 1919 euid 0: start at Mon Dec  9 21:13:40 2002
dump: version 2.4.3
running: /usr/sbin/ufsdump: dump 0Ssf 1048576 - /dev/rdsk/c1t0d0s7

There is always a killpgrp, sendsize, selfcheck, amandad, amtrmidx produced.
The sendsize has the following:
sendsize: debug 1 pid 24318 ruid 1919 euid 1919: start at Mon Dec  9
21:24:19 2002
sendsize: version 2.4.3
sendsize[24320]: time 0.005: calculating for amname '/dev/dsk/c1t0d0s7',
dirname '/h
ome', spindle -1
sendsize[24320]: time 0.005: getting size via dump for /dev/dsk/c1t0d0s7
level 0
sendsize[24320]: time 0.007: calculating for device '/dev/rdsk/c1t0d0s7'
with 'ufs'
sendsize[24320]: time 0.007: running "/usr/sbin/ufsdump 0Ssf 1048576 -
/dev/rdsk/c1t
0d0s7"
sendsize[24318]: time 0.024: waiting for any estimate child
sendsize[24320]: time 0.025: running /usr/local/libexec/killpgrp
sendsize[24320]: time 5.368: 1692931072
sendsize[24320]: time 5.369: .....
sendsize[24320]: estimate time for /dev/dsk/c1t0d0s7 level 0: 5.361
sendsize[24320]: estimate size for /dev/dsk/c1t0d0s7 level 0: 1653253 KB
sendsize[24320]: time 5.369: asking killpgrp to terminate
sendsize[24320]: time 6.369: done with amname '/dev/dsk/c1t0d0s7', dirname
'/home',
spindle -1
sendsize[24318]: time 6.370: child 24320 terminated normally
sendsize: time 6.370: pid 24318 finish time Mon Dec  9 21:24:26 2002

All the other logs appear to be normal.

Anyway hope someone can help

regards

Paul Stephens