Jean-Louis Martineau | 15 May 15:16
Favicon

Re: amvault segmentation fault

Dennis,

The attached patch fix the crash.

You have log corruption because of amdump and amvault running 
simultaneously.
Jean-Louis

On 05/15/2012 08:20 AM, Dennis Benndorf wrote:
> Am 15.05.2012 14:14, schrieb Jean-Louis Martineau:
>> On 05/15/2012 07:54 AM, Dennis Benndorf wrote:
>>>
>>> Log file /usr/local/etc/amanda/daily/catalog/log.20120506180002.0 
>>> stamped 20120506180002, expecting 20120505180458!
>> I never saw amanda print this message.
>
> This happens when amvault starts and amdump is not finished, so I 
> think they both write in the same log-file. But this has happened a 
> few times ago, mostly when our monthly backups take more than 24hours.
>
>
>> Can you run gdb again and get the value of the second argument to 
>> g_string_chunk_insert_const.
> How do I do that?
>
>
>>
>> Send me /usr/local/etc/amanda/daily/catalog/log.20120506180002.0 and 
>> /usr/local/etc/amanda/daily/catalog/log.20120505180458.0
>>
(Continue reading)

Dennis Benndorf | 14 May 14:26

amvault segmentation fault

Hello @all,

after updating our backup server to Ubuntu Server 12.04 and recompiling 
amanda 3.3.1 amvault fails with segmentation fault everytime it is 
started. Might this have something to do with the removal of 
g_queue_free_full, which was needed to compile it? Which logfiles do you 
need?

Best regards,
Dennis

gene heskett | 13 May 04:48
Favicon

The canary has died, again.

It turns out I have been running on 4.0.0alpha-4692 for a while, all the
newer ones have a broken build:

libtool: compile:  gcc -DHAVE_CONFIG_H -I. -I../config -I../gnulib -fno-strict-aliasing
-D_GNU_SOURCE -
pthread -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -Wall -Wextra -Wparentheses -Wdeclaration-
after-statement -Wmissing-prototypes -Wstrict-prototypes -Wmissing-declarations -Wformat
-Wformat-security 
-Wsign-compare -Wfloat-equal -Wold-style-definition -Wno-strict-aliasing -Wno-unknown-pragmas
-g -O2 -fno-
strict-aliasing -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -MT util.lo -MD -MP -MF .deps/util.Tpo
-c util.c  
-fPIC -DPIC -o .libs/util.o
util.c: In function 'portable_readdir':
util.c:1150:9: error: #pragma GCC diagnostic not allowed inside functions
util.c:1151:9: error: #pragma GCC diagnostic not allowed inside functions
util.c:1152:5: warning: missing initializer
util.c:1152:5: warning: (near initialization for 'mutex.unused')
util.c:1153:9: error: #pragma GCC diagnostic not allowed inside functions
make[3]: *** [util.lo] Error 1
make[3]: Leaving directory `/home/amanda/amanda-4.0.0alpha.svn.4700/common-src'
make[2]: *** [all] Error 2
make[2]: Leaving directory `/home/amanda/amanda-4.0.0alpha.svn.4700/common-src'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/amanda/amanda-4.0.0alpha.svn.4700'
make: *** [all] Error 2

Sorry to disturb the party. I was rather enjoying it myself. ;-)

Cheers, Gene
(Continue reading)

Sven Rudolph | 9 May 17:43
Picon

Specified file list matches multiple dumps

Hello,

I have a problem with amrecover:

  amrecover> add cpio.ry
  Added file /cpio.ry
  amrecover> list
  TAPE Set1-2011-18:1;Set1K-2011-13:15,16;Set1K-2011-14:1,2,3,4,5,6,7,8,9,10 LEVEL
   0 DATE 2012-01-01-03-08-50
          /cpio.ry
  amrecover> extract

  Extracting files using tape drive changer on host uxrs74.
  The following tapes are needed: Set1-2011-18 Set1K-2011-13 Set1K-2011-14

  Extracting files using tape drive changer on host uxrs74.
  Load tape Set1-2011-18 now
  Continue [?/Y/n/s/d]? y
  Specified file list matches multiple dumps; cannot continue recovery
  Got no header and data from server, check in amidxtaped.*.debug and amandad.*.debug files on server
  amrecover> 

Set1-2011-18 is the original tape, the DLE dump resides on this tape
only. Set1K-2011-13 and Set1K-2011-14 are the copies made with
amvault. I had to enable splitting in order to use amvault, so the
dump is split in several 40G-sized parts. And some of these parts are
on one tape and the other on the next tape.

I didn't split in fixed-size parts before, so this might be a cause of
the problem too.
(Continue reading)

Greg Copeland | 9 May 14:01
Picon

Adding/removing Disks?

I have a need to add a new disk and remove an existing disk. Can this be safely done by simply adding and removing the definitions in the configuration?

Thanks,

--
Greg Copeland <Greg <at> CopelandConsulting.Net>
Copeland Computer Consulting
Paul Crittenden | 1 May 18:52
Favicon

Failed tapes

What is the proper way of removing a failed tape from the tape rotation?

 

Paul Crittenden

Computer System Manager

Simpson College

Indianola, IA

direct: 515-961-1680

www.simpson.edu

 

Kervin L. Pierre | 1 May 17:41
Favicon
Gravatar

what is the proper way to delete a particlur backup ( slot )

What is the proper way to delete a particular backup slot and free the space it's holding?

I've tried 'amadmin DailySet1 delete ...' to remove 'slot1', but the files are still there in 'slot1'.  

Is it ok to 'rm -rf' in the slot folder?  Are these files referenced anywhere else?  I don't want to accidently
corrupt the install.

I was hoping there was a way to remove these files in a backup type independent way.  Eg.  I'd prefer not to have
to log into AWS to delete S3 backups, etc.

Best regards,
Kervin

Adevsoft Inc
Business Software Development
http://adevsoft.com/

Kervin L. Pierre | 1 May 02:07
Favicon
Gravatar

Device type "s3" is not known

After building from trunk to fix a few other issues, I started configuring Amazon S3 support.  

But amlabel reported...
"Device type "s3" is not known"

Similar to this issue...
https://forums.zmanda.com/archive/index.php/t-2581.html

To fix this I did...

make distclean
yum install libcurl-devel perl-WWW-Curl
./configure --enable-s3-device
make

I believe part of the underlying issue is that the autoconf tests do not seem to warn about missing headers
like libcurl, etc.

Best regards,
Kervin

Adevsoft Inc
Business Software Development
http://adevsoft.com/

Kervin L. Pierre | 29 Apr 20:20
Favicon
Gravatar

amrecover fails with "TypeError in method 'debug'..."

Hi,

I'm following the simple recovery example from the wiki...
http://wiki.zmanda.com/index.php/GSWA/Recovering_Files

Worked though all the issues but now I'm at the 'extract' and I'm getting an error...

amrecover - amidxtaped closed the connection
amrecover - can't talk to tape server: (null)

Further inspection of /var/log/messages gives...

amidxtaped[2784]: TypeError in method 'debug', argument 1 of type 'char *'

I've found someone had run into this issue before 'June 2011'...
http://archives.zmanda.com/amanda-archives/viewtopic.php?t=6929&sid=f2141406dc79480edd30a1200443e9a6

They mentioned patching a Perl script but I haven't found the one they're referring to.

Does anyone have instructions on how to fix or workaround this issue?

I'm using 3.3.1 built from source on an update to date Amazon Linux install.

Best regards,
Kervin

Adevsoft Inc
Business Software Development
http://adevsoft.com/

Greg Copeland | 26 Apr 18:17
Picon

Re: amrecover failure [RESOLVED]

On Thu, 2012-04-26 at 09:54 -0600, Steven Backus wrote:

Thanks to Steven Backus and Jean-Louis for the win. This excerpt was sent to me by Steven and did fix my issue.

I manually applied it to the perl module from my source build directory and then simply copied it into production. Of course, I made backups before I changed anything. I already had the required swig version installed.

After installation, the restore was fast and painless.

This mailing list and those who contribute are by far one of Amanda's greatest strengths.

Cheers,


1. Install swig-1.3.40. 2. Apply this patch, but not to Header.swg, to Header.pm diff --git a/perl/Amanda/Header.swg b/perl/Amanda/Header.swg index a7d391b..28cb4e8 100644 --- a/perl/Amanda/Header.swg +++ b/perl/Amanda/Header.swg <at> <at> -151,7 +151,11 <at> <at> sub from_string { sub get_dle { my $self = shift; - return Amanda::Header::HeaderXML->new($self->{'dle_str'}); + if ($self->{'dle_str'}) { + return Amanda::Header::HeaderXML->new($self->{'dle_str'}); + } else { + return undef; + } } Steve


--
Greg Copeland <Greg <at> CopelandConsulting.Net>
Copeland Computer Consulting
Greg Copeland | 26 Apr 17:43
Picon

amrecover failure

Trying to recover a file via amrecover. I set the host and disk. I add the file in question. I confirm the tape is in the drive. I tell it to extract.

I get:
Got no header and data from server, check in amidxtaped.*.debug and amandad.*.debug files on server

Any help? I have no idea where to begin. I can't seem to find the debug files in question. I wonder if this might be because in order to get Amanda to run in the first place, per my previous recovery, I had to comment out debugging output.

Two recovery attempts and two times Amanda has failed to deliver. I have used Amanda years ago and only recently started using it again. I never had issues in the past. Something seems to have gone terribly wrong with Amanda since days gone by.

Help,


--
Greg Copeland <Greg <at> CopelandConsulting.Net>
Copeland Computer Consulting

Gmane