Amanda-Users

sendbackup leaves processes running

2007-06-20 11:24:17
Subject: sendbackup leaves processes running
From: Jean-Francois Malouin <Jean-Francois.Malouin AT bic.mni.mcgill DOT ca>
To: AMANDA users <amanda-users AT amanda DOT org>
Date: Wed, 20 Jun 2007 11:19:17 -0400
Hi,

I've seen this many times with 2.5.2p1 and previous 2.5.x snapshots
and I have reported it before but I got no replies. Server is running
irix-6.5.x

A DLE fails to make it to tape (in this particular case it hit EOT and
amanda doesn't seem to a retry btw) and gtar is left running. Output
from ps shows:

root   24578406 1  0 09:33:15 ? 68:08 gtar --create --file - --directory 
/data/cortfmri/cortfmri1/jens/mavan --one-fi

See the attached sendbackup and runtar debug file. I'm also wondering
why this DLE wasn't retried after hitting EOT. Is it because it was
directly sent to tape? Or some thing times out? The amdump log shows:


driver: result time 40461.104 from dumper0: DONE 00-00078 10 10 223 "[sec 
222.566 kb 10 kps 0.0 orig-kb 10]"
taper: reader-side: got label stk_40-conf4-000029 filenum 76
driver: result time 40461.589 from taper: DONE 00-00078 stk_40-conf4-000029 76 
"[sec 223.154 kb 32 kps 0.1 {wr: writers 2 rdwait 220.792 wrwait 0.482 filemark 
1.878}]"
driver: dumping yorick:cortfmri1_jens_mavan_C directly to tape
driver: send-cmd time 40461.610 to taper: PORT-WRITE 00-00079 yorick 
ffffffff9ffeffffffff00 cortfmri1_jens_mavan_C 0 20070619 0 NULL 10240
driver: result time 40461.620 from taper: PORT 17581
driver: send-cmd time 40461.621 to dumper0: PORT-DUMP 00-00079 17581 yorick 
ffffffff9ffeffffffff00 cortfmri1_jens_mavan_C 
/data/cortfmri/cortfmri1/jens/mavan 0 1970:1:1:0:0:0 GNUTAR X X X 
|;auth=bsdtcp;index;include-file=./C*;
driver: state time 40461.621 free kps: 1013402 space: 0 taper: writing 
idle-dumpers: 11 qlen tapeq: 0 runq: 0 roomq: 0 wakeup: 0 driver-idle: not-idle
driver: interface-state time 40461.621 if default: free 1012002 if local: free 
1000 if le0: free 400
driver: hdisk-state time 40461.621 hdisk 0: free 0 dumpers 0
send request:
----
SERVICE sendbackup
OPTIONS features=ffffffff9ffeffffffff00;hostname=yorick;config=stk_80-conf4;
GNUTAR cortfmri1_jens_mavan_C /data/cortfmri/cortfmri1/jens/mavan 0 
1970:1:1:0:0:0 OPTIONS |;auth=bsdtcp;index;include-file=./C*;

----

got response:
----
CONNECT DATA 499999 MESG 499998 INDEX 499997
OPTIONS features=ffffffff9ffeffffffff00;

----

taper: writing end marker. [stk_40-conf4-000029 ERR kb 104075648 fm 77]
changer: opening pipe to: /opt/amanda/amanda4/libexec/chg-zd-mtx -info
changer: opening pipe to: /opt/amanda/amanda4/libexec/chg-zd-mtx -search 
stk_40-conf4-000030
taper: slot: 52 wrote label `stk_40-conf4-000030' date `20070619'
dumper: kill index command
driver: result time 41176.593 from dumper0: FAILED 00-00079 "[data write: 
Broken pipe]"
driver: result time 41176.593 from taper: TRY-AGAIN 00-00079 "[writing file: No 
space left on device]"
driver: state time 41176.679 free kps: 1025400 space: 0 taper: idle 
idle-dumpers : 12 qlen tapeq: 0 runq: 0 roomq: 0 wakeup: 0 driver-idle: not-idle
driver: interface-state time 41176.679 if default: free 1024000 if local: free 
1 000 if le0: free 400
driver: hdisk-state time 41176.679 hdisk 0: free 0 dumpers 0
driver: QUITTING time 41176.680 telling children to quit
driver: send-cmd time 41176.680 to dumper0: QUIT
driver: send-cmd time 41176.680 to dumper1: QUIT
driver: send-cmd time 41176.680 to dumper2: QUIT
driver: send-cmd time 41176.680 to dumper3: QUIT
driver: send-cmd time 41176.680 to dumper4: QUIT
driver: send-cmd time 41176.680 to dumper5: QUIT
driver: send-cmd time 41176.680 to dumper6: QUIT
driver: send-cmd time 41176.680 to dumper7: QUIT
driver: send-cmd time 41176.680 to dumper8: QUIT
driver: send-cmd time 41176.681 to dumper9: QUIT
driver: send-cmd time 41176.681 to dumper10: QUIT
driver: send-cmd time 41176.681 to dumper11: QUIT
taper: DONE [idle wait: 861.394 secs]
taper: writing end marker. [stk_40-conf4-000030 OK kb 0 fm 0]
driver: FINISHED time 41183.753
amdump: end at Wed Jun 20 09:41:37 EDT 2007

In any case this is triply bad as the DLE wasn't dumped,
a tape got wasted and some processes were left behind...

jf
-- 
<° ><

Attachment: runtar.20070620093315.debug
Description: Text document

Attachment: sendbackup.20070620092935.debug
Description: Text document

<Prev in Thread] Current Thread [Next in Thread>
  • sendbackup leaves processes running, Jean-Francois Malouin <=