Amanda-Users

Re: 2.4.5: dumping directly to tape failure: no retry?

2005-11-03 11:44:53
Subject: Re: 2.4.5: dumping directly to tape failure: no retry?
From: Jean-Francois Malouin <Jean-Francois.Malouin AT bic.mni.mcgill DOT ca>
To: AMANDA users <amanda-users AT amanda DOT org>, Jean-Louis Martineau <martinea AT iro.umontreal DOT ca>
Date: Thu, 3 Nov 2005 11:27:06 -0500
* Jean-Francois Malouin <malin AT bic.mni.mcgill DOT ca> [20051102 07:58]:
> * Jean-Louis Martineau <martinea AT iro.umontreal DOT ca> [20051101 19:56]:
> > Hi Jean-Francois,
> > 
> > Thanks for the good report.
> > 
> > The problem is the result of the dumper, which is FAILED, in that
> > case the dump is not retried.
> > 
> > Could you try this patch.
> 
> I just applied the patch and will let know how it goes.

Looks like that patch did the work! Amanda retried a full backup of
the filesystem /data/mril/mril2 that failed and luckily enough the
same situation arised as yesterday when amanda hit EOT while dumping
directly to tape. But this time it retried and dumped it successfully.

----
amdump log excerpts (sorry for the long lines!)
----

driver: dumping yorick:/data/mril/mril2 directly to tape

    [...]

driver: send-cmd time 4867.717 to taper: PORT-WRITE 00-00031 yorick
fffffeff9ffe7f /data/mril/mril2 0 20051103
driver: send-cmd time 4867.721 to dumper0: PORT-DUMP 00-00031 16845
yorick fffffeff9ffe7f /data/mril/mril2 NODEVICE 0 1970:1:1:0:0:0 DUMP
|;bsd-auth;index;

    [...]

taper: writing end marker. [stk_40-conf2-000022 ERR kb 103840288 fm 16]
changer: opening pipe to: /opt/amanda/amanda2/libexec/chg-zd-mtx -info
changer: got exit: 0 str: 15 10 1 1
changer_query: changer return was 10 1 1
changer_query: searchable = 1
changer_find: looking for stk_40-conf2-000023 changer is searchable = 1
changer_search: stk_40-conf2-000023
changer: opening pipe to: /opt/amanda/amanda2/libexec/chg-zd-mtx -search 
stk_40-conf2-000023
changer: got exit: 0 str: 16 /hw/tape/tps12d2nrnsv
taper: slot 16: date 20050930 label stk_40-conf2-000023 (exact label match)
taper: read label `stk_40-conf2-000023' date `20050930'
taper: wrote label `stk_40-conf2-000023' date `20051103'
dumper: kill index command
driver: result time 12869.605 from dumper0: TRY-AGAIN 00-00031 ["data write: 
Broken pipe"]
driver: result time 12869.606 from taper: TRY-AGAIN 00-00031 [writing file: No 
space left on device]
driver: error time 12869.632 serial gen mismatch
driver: dumping yorick:/data/mril/mril2 directly to tape
driver: send-cmd time 12869.633 to taper: PORT-WRITE 00-00032 yorick 
fffffeff9ffe7f /data/mril/mril2 0 20051103
driver: send-cmd time 12869.637 to dumper0: PORT-DUMP 00-00032 16655 yorick 
fffffeff9ffe7f /data/mril/mril2 NODEVICE 0 1970:1:1:0:0:0 DUMP |;bsd-auth;index;
driver: result time 21175.492 from dumper0: DONE 00-00032 9630451696304515 8306 
[sec 8305.590 kb 96304515 kps 11595.1 orig-kb 96304516]
taper: reader-side: got label stk_40-conf2-000023 filenum 1
driver: result time 21175.550 from taper: DONE 00-00032 stk_40-conf2-000023 1 
[sec 8305.906 kb 96304516 kps 11594.7 {wr: writers 3009518 rdwait 2319.975 
wrwait 5796.334 filemark 2.688}]


Thanks again. This list is fantastic!
jf

> 
> Thanks a bunch,
> jf
> 
> > 
> > 
> > On Tue, Nov 01, 2005 at 11:37:02AM -0500, Jean-Francois Malouin wrote:
> > > dumper: driver: result time 23806.848 from dumper0: FAILED 00-00031
> > 
> > -- 
> > Jean-Louis Martineau             email: martineau AT iro.umontreal DOT ca 
> > Département IRO, Université de Montréal
> > C.P. 6128, Succ. CENTRE-VILLE    Tel: (514) 343-6111 ext. 1-3529
> > Montréal, Canada, H3C 3J7        Fax: (514) 343-5834
> 
> > --- amanda-2.4.5p1.orig/server-src/dumper.c 2005-09-20 20:11:38.000000000 
> > -0400
> > +++ amanda-2.4.5p1.new/server-src/dumper.c  2005-11-01 19:42:29.000000000 
> > -0500
> > @@ -1387,8 +1387,7 @@ int mesgfd, datafd, indexfd, outfd;
> >             rc = 2;
> >             goto failed;
> >         }
> > -       if(update_dataptr(&outfd, size1)) {
> > -           rc = 2;
> > +       if((rc=update_dataptr(&outfd, size1))) {
> >             goto failed;
> >         }
> >         if(size1 == 0) {
> 
> 

-- 
<° ><