Paul Bijnens wrote:
H.Caglar Bilir wrote:
the problem is, how can i trace this error. i looked at the x machines' /tmp/amanda directory and found a core file dated as just after sendsize file and before amandad file. the output on amandad file is as follows:
I have no recent experience with AIX. Try "file core". Some implementations of "file" indicate which program crashed.
"strings core" can help too.
####################### Amanda 2.4 REQ HANDLE 023-2005C048 SEQ 1057772088 SECURITY USER dumpuser SERVICE sendsize OPTIONS maxdumps=1;hostname=x; GNUTAR hd2 0 1970:1:1:0:0:0 -1 DUMP hd1 0 1970:1:1:0:0:0 -1 --------
amandad: time 0.000: sending ack: ---- Amanda 2.4 ACK HANDLE 023-2005C048 SEQ 1057772088 ----
amandad: time 0.012: bsd security: remote host y.cc.metu.edu.tr user dumpuse
r local user dumpuser
amandad: time 0.012: amandahosts security check passed
amandad: time 0.012: running service "/usr/local/libexec/sendsize"
amandad: time 0.190: sending REP packet:
----
Amanda 2.4 REP HANDLE 023-2005C048 SEQ 1057772088
At this position, I would expect to see the actual sizes. The reply packet seems empty. So the core is probably from sendsize. There should be a file "sendsize.aboutsamedatestamp.debug" too. Any error message in that file, (maybe just before dumping core)?
both sendsize and selfcheck programs generate core file(after some hand tests - and core files are renamed)
if i look to the actual backup time : core file is because sendsize. the related sendsize file is as follows:
>less sendsize.20030709203046.debug
sendsize: debug 1 pid 528590 ruid 211 euid 211: start at Wed Jul 9 20:30:46 200
3
sendsize: version 2.4.4p1
any ideas?
