Quantcast

Re: [Mondo-devel] mondoarchive 2.2.9.2-1 fails with "buffer overflow detected"

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] mondoarchive 2.2.9.2-1 fails with "buffer overflow detected"

Tom Metro
Bruno Cornec wrote:
> BTW I fixed the problem I had with the Ubuntu 9.04 x86_64 VM and
> generated package for you to try.

Thanks. Are you aware that there is a version skew between your packages
and the ones in the Ubuntu repositories, which have a "1:" prefix?

# aptitude-availableversions mondo
mondo:
   Installed: (none)
   Candidate: 1:2.2.7-2
   Version table:
      1:2.2.7-2 0
         500 http://us.archive.ubuntu.com jaunty/universe Packages
      2.2.9.2-1 0
         500 ftp://ftp.mondorescue.org 9.04/contrib Packages

# aptitude-availableversions mindi
mindi:
   Installed: (none)
   Candidate: 1:2.0.4-1ubuntu1
   Version table:
      1:2.0.4-1ubuntu1 0
         500 http://us.archive.ubuntu.com jaunty/universe Packages
         100 /var/lib/dpkg/status
      2.0.7.2-1 0
         500 ftp://ftp.mondorescue.org 9.04/contrib Packages

To get the versions from ftp.mondorescue.org I had to do:

# aptitude install mondo=2.2.9.2-1 mindi=2.0.7.2-1


Before mindi was failing. Now mondoarchive is failing before it gets
that far. It dies just after specifying the path to include in the
backup using the interactive UI. Here are the interesting bits from the log:

                [Main]
libmondo-devices.c->interactively_obtain_media_parameters_from_user#2085:
media type = default
        [Main] libmondo-tools.c->setup_tmpdir#715: Purging old tmpdir
//mondo.tmp.A6WIM6
        [Main] libmondo-devices.c->sensibly_set_tmpdir_and_scratchdir#2808:
bkpinfo->tmpdir is being set to //mondo.tmp
.3fmOVP
        [Main] libmondo-devices.c->sensibly_set_tmpdir_and_scratchdir#2812:
bkpinfo->scratchdir is being set to //mondo
.scratch.16174

        [Main] libmondo-devices.c->find_dvd_device#924: I think DVD is at /dev/scd0
[Main]
libmondo-devices.c->interactively_obtain_media_parameters_from_user#2125:
Setting to DVD defaults
[Main]
libmondo-devices.c->interactively_obtain_media_parameters_from_user#2169:
bkpinfo->media_device = /dev/scd0
[Main]
libmondo-devices.c->interactively_obtain_media_parameters_from_user#2173:
bkpinfo->media_device = /dev/scd0
        [Main]
libmondo-devices.c->interactively_obtain_media_parameters_from_user#2183:
DVD device found at /dev/scd0
        [Main] libmondo-devices.c->which_boot_loader#3031: list_drives_cmd =
parted2fdisk -l 2>/dev/null | grep "/dev/.
*:" | tr -s ':' ' ' | tr -s ' ' '
' | grep /dev/; echo /dev/sda
        [Main] libmondo-devices.c->which_boot_loader#3044: looking at drive
/dev/sda's MBR
        [Main] libmondo-devices.c->which_boot_loader#3059: 1 grubs and 0 lilos

SIGABRT signal received from OS
Abort - probably failed assertion. I'm sleeping for a few seconds so you
can read the message.
[Main] newt-specific.c->fatal_error#308: Fatal error received -
'MondoRescue is terminating in response to a signal fro
m the OS'
                [Main] newt-specific.c->fatal_error#326: OK, I think I'm the main PID.
        [Main] newt-specific.c->fatal_error#334: I'm going to do some cleaning
up now.
                        [Main] newt-specific.c->fatal_error#335: killall mindi 2> /dev/null
running: ps auxww | grep " /mondo/do-not " | awk '{print $2;}' | grep -v
"grep" > //mondo.tmp.3fmOVP/mondo-run-prog-thi
ng.tmp 2> //mondo.tmp.3fmOVP/mondo-run-prog-thing.err
--------------------------------start of output-----------------------------
9937
9940
--------------------------------end of output------------------------------
...ran just fine. :-)
running: kill `ps auxww | grep " /mondo/do-not " | awk '{print $2;}' |
grep -v "grep"` > //mondo.tmp.3fmOVP/mondo-run-p
rog-thing.tmp 2> //mondo.tmp.3fmOVP/mondo-run-prog-thing.err
--------------------------------start of output-----------------------------
--------------------------------end of output------------------------------
...ran with res=15
running: ps auxww | grep " tmp.mondo " | awk '{print $2;}' | grep -v
"grep" > //mondo.tmp.3fmOVP/mondo-run-prog-thing.t
mp 2> //mondo.tmp.3fmOVP/mondo-run-prog-thing.err
--------------------------------start of output-----------------------------
9959
9961
--------------------------------end of output------------------------------
...ran just fine. :-)
running: kill `ps auxww | grep " tmp.mondo " | awk '{print $2;}' | grep
-v "grep"` > //mondo.tmp.3fmOVP/mondo-run-prog-
thing.tmp 2> //mondo.tmp.3fmOVP/mondo-run-prog-thing.err
--------------------------------start of output-----------------------------
--------------------------------end of output------------------------------
...ran with res=15
running: ps auxww | grep " ntfsclone " | awk '{print $2;}' | grep -v
"grep" > //mondo.tmp.3fmOVP/mondo-run-prog-thing.t
mp 2> //mondo.tmp.3fmOVP/mondo-run-prog-thing.err
--------------------------------start of output-----------------------------
9982
9986
--------------------------------end of output------------------------------
...ran just fine. :-)
running: kill `ps auxww | grep " ntfsclone " | awk '{print $2;}' | grep
-v "grep"` > //mondo.tmp.3fmOVP/mondo-run-prog-
thing.tmp 2> //mondo.tmp.3fmOVP/mondo-run-prog-thing.err
--------------------------------start of output-----------------------------
--------------------------------end of output------------------------------
...ran with res=15
-------FATAL ERROR---------
MondoRescue is terminating in response to a signal from the OS
running: umount /mnt/cdrom > //mondo.tmp.3fmOVP/mondo-run-prog-thing.tmp
2> //mondo.tmp.3fmOVP/mondo-run-prog-thing.err
--------------------------------start of output-----------------------------
umount: /mnt/cdrom: not mounted
--------------------------------end of output------------------------------
...ran with res=256
[Main] newt-specific.c->finish#422: Calling newtSuspend()


Also note how the temp files are all using a path like //mondo.tmp*,
with an apparently missing tmp directory name. Setting TMP environment
variable seemed to fix that, but it shouldn't be necessary to do that.

I re-installed the locally built mondo_2.2.9.2-1_amd64.deb, while
keeping the mindi from ftp.mondorescue.org, and tried again.

Same failure. I notice I'm seeing an error on the screen that isn't
making it into the log:

*** buffer overflow detected ***: mondoarchive terminated
======= Backtrace: =========
/lib/libc.so.6(__fortify_fail+0x37)[0x7f5e245cb2c7]
/lib/libc.so.6[0x7f5e245c9170]
mondoarchive[0x42ec3d]
mondoarchive[0x414b42]
mondoarchive[0x403c67]
/lib/libc.so.6(__libc_start_main+0xe6)[0x7f5e244ea5a6]
mondoarchive[0x403059]
======= Memory map: ========
00400000-00451000 r-xp 00000000 08:01 10945416
  /usr/sbin/mondoarchive
00650000-00651000 r--p 00050000 08:01 10945416
  /usr/sbin/mondoarchive
00651000-00653000 rw-p 00051000 08:01 10945416
  /usr/sbin/mondoarchivel rights reversed.
00653000-00657000 rw-p 00653000 00:00 0
015f6000-0167c000 rw-p 015f6000 00:00 0
  [heap]
7f5e238d1000-7f5e238e7000 r-xp 00000000 08:01 1769533
  /lib/libgcc_s.so.1
7f5e238e7000-7f5e23ae7000 ---p 00016000 08:01 1769533
  /lib/libgcc_s.so.1
7f5e23ae7000-7f5e23ae8000 r--p 00016000 08:01 1769533
  /lib/libgcc_s.so.1
7f5e23ae8000-7f5e23ae9000 rw-p 00017000 08:01 1769533
  /lib/libgcc_s.so.1
7f5e23ae9000-7f5e23af7000 r-xp 00000000 08:01 7800647
  /usr/lib/libfribidi.so.0.0.0
7f5e23af7000-7f5e23cf7000 ---p 0000e000 08:01 7800647
  /usr/lib/libfribidi.so.0.0.0
7f5e23cf7000-7f5e23cf8000 rw-p 0000e000 08:01 7800647
  /usr/lib/libfribidi.so.0.0.0
7f5e23cf8000-7f5e23d7c000 r-xp 00000000 08:01 1769869
  /lib/libm-2.9.so
7f5e23d7c000-7f5e23f7b000 ---p 00084000 08:01 1769869
  /lib/libm-2.9.so
7f5e23f7b000-7f5e23f7c000 r--p 00083000 08:01 1769869
  /lib/libm-2.9.so
7f5e23f7c000-7f5e23f7d000 rw-p 00084000 08:01 1769869
  /lib/libm-2.9.so
7f5e23f7d000-7f5e23f7f000 r-xp 00000000 08:01 1769868
  /lib/libdl-2.9.so
7f5e23f7f000-7f5e2417f000 ---p 00002000 08:01 1769868
  /lib/libdl-2.9.so
7f5e2417f000-7f5e24180000 r--p 00002000 08:01 1769868
  /lib/libdl-2.9.so
7f5e24180000-7f5e24181000 rw-p 00003000 08:01 1769868
  /lib/libdl-2.9.so
7f5e24181000-7f5e24251000 r-xp 00000000 08:01 1769924
  /lib/libslang.so.2.1.3
7f5e24251000-7f5e24451000 ---p 000d0000 08:01 1769924
  /lib/libslang.so.2.1.3
7f5e24451000-7f5e24455000 r--p 000d0000 08:01 1769924
  /lib/libslang.so.2.1.3
7f5e24455000-7f5e2446c000 rw-p 000d4000 08:01 1769924
  /lib/libslang.so.2.1.3
7f5e2446c000-7f5e244cc000 rw-p 7f5e2446c000 00:00 0
7f5e244cc000-7f5e24634000 r-xp 00000000 08:01 1769865
  /lib/libc-2.9.so
7f5e24634000-7f5e24834000 ---p 00168000 08:01 1769865
  /lib/libc-2.9.so
7f5e24834000-7f5e24838000 r--p 00168000 08:01 1769865
  /lib/libc-2.9.so
7f5e24838000-7f5e24839000 rw-p 0016c000 08:01 1769865
  /lib/libc-2.9.so
7f5e24839000-7f5e2483e000 rw-p 7f5e24839000 00:00 0
7f5e2483e000-7f5e24855000 r-xp 00000000 08:01 1769884
  /lib/libpthread-2.9.so
7f5e24855000-7f5e24a54000 ---p 00017000 08:01 1769884
  /lib/libpthread-2.9.so
7f5e24a54000-7f5e24a55000 r--p 00016000 08:01 1769884
  /lib/libpthread-2.9.so
7f5e24a55000-7f5e24a56000 rw-p 00017000 08:01 1769884
  /lib/libpthread-2.9.so
7f5e24a56000-7f5e24a5a000 rw-p 7f5e24a56000 00:00 0
7f5e24a5a000-7f5e24a6f000 r-xp 00000000 08:01 7800686
  /usr/lib/libnewt.so.0.52.2
7f5e24a6f000-7f5e24c6f000 ---p 00015000 08:01 7800686
  /usr/lib/libnewt.so.0.52.2
7f5e24c6f000-7f5e24c70000 r--p 00015000 08:01 7800686
  /usr/lib/libnewt.so.0.52.2
7f5e24c70000-7f5e24c71000 rw-p 00016000 08:01 7800686
  /usr/lib/libnewt.so.0.52.2
7f5e24c71000-7f5e24c91000 r-xp 00000000 08:01 1769859
  /lib/ld-2.9.so
7f5e24d3b000-7f5e24d7a000 r--p 00000000 08:01 7831690
  /usr/lib/locale/en_US.utf8/LC_CTYPE
7f5e24d7a000-7f5e24e65000 r--p 00000000 08:01 7831692
  /usr/lib/locale/en_US.utf8/LC_COLLATE
7f5e24e65000-7f5e24e69000 rw-p 7f5e24e65000 00:00 0
7f5e24e7b000-7f5e24e7c000 r--p 00000000 08:01 7832503
  /usr/lib/locale/en_US.utf8/LC_NUMERIC
7f5e24e7c000-7f5e24e7d000 r--p 00000000 08:01 7832251
  /usr/lib/locale/en_US.utf8/LC_TIME
7f5e24e7d000-7f5e24e7e000 r--p 00000000 08:01 7832252
  /usr/lib/locale/en_US.utf8/LC_MONETARY
7f5e24e7e000-7f5e24e7f000 r--p 00000000 08:01 7839766
  /usr/lib/locale/en_US.utf8/LC_MESSAGES/SYS_LC_MESSAGES
7f5e24e7f000-7f5e24e80000 r--p 00000000 08:01 7832504
  /usr/lib/locale/en_US.utf8/LC_PAPER
7f5e24e80000-7f5e24e81000 r--p 00000000 08:01 7832502
  /usr/lib/locale/en_US.utf8/LC_NAME
7f5e24e81000-7f5e24e82000 r--p 00000000 08:01 7832253
  /usr/lib/locale/en_US.utf8/LC_ADDRESS
7f5e24e82000-7f5e24e83000 r--p 00000000 08:01 7832254
  /usr/lib/locale/en_US.utf8/LC_TELEPHONE
7f5e24e83000-7f5e24e84000 r--p 00000000 08:01 7832393
  /usr/lib/locale/en_US.utf8/LC_MEASUREMENT
7f5e24e84000-7f5e24e8b000 r--s 00000000 08:01 7816807
  /usr/lib/gconv/gconv-modules.cache
7f5e24e8b000-7f5e24e8c000 r--p 00000000 08:01 7832395
  /usr/lib/locale/en_US.utf8/LC_IDENTIFICATION
Fatal error... MondoRescue is terminating in response to a signal from
the OS
rw-p 7f5e24e8c000 00:00 0 7f5e24e90000-
e91000-7f5e24e92000 rw-p 00020000 08:01
ea000 00:00 0                      [stao]
SIGABRT signal received from OS

ffffffffff600000-ffffffffff601000 r-xAbort - probably failed assertion.
I'm sleeping for a few seconds so you can re
---FATALERROR--- MondoRescue is terminating in response to a signal from
the OS


  -Tom

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] mondoarchive 2.2.9.2-1 fails with "buffer overflow detected"

Bruno Cornec-2
Tom Metro said on Sat, Apr 10, 2010 at 10:21:44PM -0400:

> Bruno Cornec wrote:
> > BTW I fixed the problem I had with the Ubuntu 9.04 x86_64 VM and
> > generated package for you to try.
>
> Thanks. Are you aware that there is a version skew between your packages
> and the ones in the Ubuntu repositories, which have a "1:" prefix?

Sort of. It should be now well documented in the FAQ of the wiki
(http://trac.mondorescue.org/wiki/FAQ#Q11DoesmondoworkwithDebianUbuntudistributions)

> [Main] libmondo-devices.c->which_boot_loader#3044: looking at drive
> /dev/sda's MBR
> [Main] libmondo-devices.c->which_boot_loader#3059: 1 grubs and 0 lilos
>
> SIGABRT signal received from OS
> Abort - probably failed assertion. I'm sleeping for a few seconds so you
> can read the message.

Could you try again with 2.2.9.3 just out (to be announced here) and
report again please ?

> Also note how the temp files are all using a path like //mondo.tmp*,
> with an apparently missing tmp directory name. Setting TMP environment
> variable seemed to fix that, but it shouldn't be necessary to do that.

Please create a trac entry for that.

> ffffffffff600000-ffffffffff601000 r-xAbort - probably failed assertion.

If you are able to rebuild packages, you may want to try this recipe at
http://trac.mondorescue.org/wiki/TroubleShooting#Trouble-Shootingmondo
if the latest version doesn't fix it for you.

Best regards,
Bruno.
--
Open Source & Linux Profession Lead EMEA           / http://opensource.hp.com
HP/Intel/Red Hat Open Source Solutions Initiative  / http://www.hpintelco.net
http://www.HyPer-Linux.org  http://mondorescue.org http://project-builder.org
La musique ancienne?  http://www.musique-ancienne.org http://www.medieval.org

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] mondoarchive 2.2.9.3 package for Ubuntu 9.04

Tom Metro
Bruno Cornec wrote:
> Could you try again with 2.2.9.3 just out (to be announced here) and
> report again please ?

I'm still seeing 2.2.9.2-1 as the latest version available from
ftp.mondorescue.org:

% sudo aptitude update
% apt-cache policy mondo
mondo:
   Installed: 1:2.2.7-2
   Candidate: 1:2.2.7-2
   Version table:
  *** 1:2.2.7-2 0
         500 http://us.archive.ubuntu.com jaunty/universe Packages
         100 /var/lib/dpkg/status
      2.2.9.2-1 0
         500 ftp://ftp.mondorescue.org 9.04/contrib Packages


  -Tom


------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] Debian/Ubuntu package version skew

Tom Metro
In reply to this post by Bruno Cornec-2
Bruno Cornec wrote:
>> Thanks. Are you aware that there is a version skew between your packages
>> and the ones in the Ubuntu repositories, which have a "1:" prefix?
>
> Sort of. It should be now well documented in the FAQ of the wiki
> (http://trac.mondorescue.org/wiki/FAQ#Q11DoesmondoworkwithDebianUbuntudistributions)

Thanks for the pointer. I missed that.

I've updated my pinning.


The FAQ says:
> If you want to use the packages made upstream for Debian/Ubuntu, you
> need to know that numbering schema are different between upstream mondo
> (2.x.y) and Debian packages (2.xy)

That doesn't accurately describe the package versions I see in Ubuntu.
Is Debian different, or is this information obsolete?

Looks like the above was true for Etch. The package doesn't exist in
Lenny (which seems strange). And Squeeze and later used the "1:" prefix,
that Ubuntu is picking up.
http://packages.debian.org/source/squeeze/mondo

Out of curiosity, what led to this mismatch?

  -Tom


------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] Debian/Ubuntu package version skew

Doug Laidlaw-2-3
On Sat, 17 Apr 2010 02:39:20 -0400
Tom Metro <[hidden email]> wrote:

> Bruno Cornec wrote:
> >> Thanks. Are you aware that there is a version skew between your
> >> packages and the ones in the Ubuntu repositories, which have a
> >> "1:" prefix?
> >
> > Sort of. It should be now well documented in the FAQ of the wiki
> > (http://trac.mondorescue.org/wiki/FAQ#Q11DoesmondoworkwithDebianUbuntudistributions)
>
> Thanks for the pointer. I missed that.
>
> I've updated my pinning.
>
>
> The FAQ says:
> > If you want to use the packages made upstream for Debian/Ubuntu, you
> > need to know that numbering schema are different between upstream
> > mondo (2.x.y) and Debian packages (2.xy)
>
> That doesn't accurately describe the package versions I see in
> Ubuntu. Is Debian different, or is this information obsolete?
>
> Looks like the above was true for Etch. The package doesn't exist in
> Lenny (which seems strange). And Squeeze and later used the "1:"
> prefix, that Ubuntu is picking up.
> http://packages.debian.org/source/squeeze/mondo
>
> Out of curiosity, what led to this mismatch?
>
>   -Tom
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Mondo-devel mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/mondo-devel
>
> !DSPAM:4bc99bfb239261216617018!
>

The same mismatch occurs with Mandriva.  As Bruno is on leave, I will
risk my neck with an explanation.  (I usually put my foot in it
somehow.)  About version 2.1.6 (from memory) an error crept in, and it
went out as 2.16. Mandriva and some other distros have changed their
package numbers to a 2.16-style series, but Bruno is still using the old
system.  I wasn't aware that there was a fix available for Debian;
AFAIK there is nothing similar for RPMs.  I have my own repository of
downloaded RPMs.  I can tell urpmi to ignore the "updates" which would
put the old version back, because e.g. 2.1.9 is before 2.16.  Hence
Bruno's comment yesterday that a 3.0.0 series would solve the whole
problem.

Doug.

------------------------------------------------------------------------------
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] Debian/Ubuntu package version skew

Doug Laidlaw
In reply to this post by Tom Metro
On Sat, 17 Apr 2010 02:39:20 -0400
Tom Metro <[hidden email]> wrote:


> The FAQ says:
> > If you want to use the packages made upstream for Debian/Ubuntu, you
> > need to know that numbering schema are different between upstream
> > mondo (2.x.y) and Debian packages (2.xy)
>
> That doesn't accurately describe the package versions I see in
> Ubuntu. Is Debian different, or is this information obsolete?
>
> Looks like the above was true for Etch. The package doesn't exist in
> Lenny (which seems strange). And Squeeze and later used the "1:"
> prefix, that Ubuntu is picking up.
> http://packages.debian.org/source/squeeze/mondo
>
> Out of curiosity, what led to this mismatch?
>
>   -Tom
>
>

(My previous reply was held because my address was wrong :(  )
The same mismatch occurs with Mandriva.  As Bruno is on leave, I will
risk my neck with an explanation.  (I usually put my foot in it
somehow.)  About version 2.1.6 (from memory) an error crept in, and it
went out as 2.16. Mandriva and some other distros have changed their
package numbers to a 2.16-style series, but Bruno is still using the old
system.  I wasn't aware that there was a fix available for Debian;
AFAIK there is nothing similar for RPMs.  I have my own repository of
recently downloaded RPMs.  I can tell urpmi to ignore the "updates"
which would put the old version back, because e.g. 2.1.9 is seen as
earlier than 2.16. Hence Bruno's comment yesterday that a 3.0.0 series
would solve the whole problem.

Even if the numbers are incorrect, they demonstrate the problem.

Doug.

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] Debian/Ubuntu package version skew

Tom Metro
Doug Laidlaw wrote:
> Tom Metro wrote:
>> Out of curiosity, what led to this mismatch?
>
> About version 2.1.6 (from memory) an error crept in, and it
> went out as 2.16.

Ah. I figured it was something like that. Thanks for the history.

  -Tom


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] mondoarchive 2.2.9.3 package for Ubuntu 9.04

Tom Metro
In reply to this post by Tom Metro
Tom Metro wrote:
> I'm still seeing 2.2.9.2-1 as the latest version available from
> ftp.mondorescue.org:

Looks like the 64-bit packages did not get built.

I'll create a package from source, if I have time to test it before
Bruno gets back.

  -Tom

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] Debian/Ubuntu package version skew

Bruno Cornec-2
In reply to this post by Tom Metro
Tom Metro said on Sat, Apr 17, 2010 at 02:39:20AM -0400:

> That doesn't accurately describe the package versions I see in Ubuntu.
> Is Debian different, or is this information obsolete?

I'll let people more versed in Debian/Ubuntu give accurate answers to
that.

> Looks like the above was true for Etch. The package doesn't exist in
> Lenny (which seems strange). And Squeeze and later used the "1:" prefix,
> that Ubuntu is picking up.
> http://packages.debian.org/source/squeeze/mondo

Well, Lenny didn't kept mondo due to bugs reported at that time (which
weren't not real ones, but anyway).

Happy to see Squeeze picking up. However, I'm not sure the 1: prefix is
a good idea (what will happen when 3.0.0 is out ?)

> Out of curiosity, what led to this mismatch?


History as Doug mentioned, and my will to have correct numbers for
versions (2.0.x instead of 2.0x which were already a problem for package
update).
Now doing it the proper way, I created another proble for some distros.
Which lead me to think renaming the next version 3.x is the Good Thing
to do ;-)

Bruno.
--
Open Source & Linux Profession Lead EMEA           / http://opensource.hp.com
HP/Intel/Red Hat Open Source Solutions Initiative  / http://www.hpintelco.net
http://www.HyPer-Linux.org  http://mondorescue.org http://project-builder.org
La musique ancienne?  http://www.musique-ancienne.org http://www.medieval.org

------------------------------------------------------------------------------
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] mondoarchive 2.2.9.3 package for Ubuntu 9.04

Bruno Cornec-2
In reply to this post by Tom Metro
Tom Metro said on Sat, Apr 17, 2010 at 03:32:18PM -0400:

> Looks like the 64-bit packages did not get built.

Yes, on 9.04 I have a pb with the x86_64 build I need to look at.
You have packages for 0.10 however.

Bruno.
--
Open Source & Linux Profession Lead EMEA           / http://opensource.hp.com
HP/Intel/Red Hat Open Source Solutions Initiative  / http://www.hpintelco.net
http://www.HyPer-Linux.org  http://mondorescue.org http://project-builder.org
La musique ancienne?  http://www.musique-ancienne.org http://www.medieval.org

------------------------------------------------------------------------------
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [Mondo-devel] mond, Kleo Bare Metal Backup, live CD

Tom Metro
Bruno Cornec wrote:
> Tom Metro said:
>> Looks like the 64-bit packages did not get built.
>
> Yes, on 9.04 I have a pb with the x86_64 build I need to look at.
> You have packages for 0.10 however.

You mean 9.10?


My objective was to use Mondo to backup a 9.04 installation to DVDs
prior to upgrading to 9.10.

After multiple tries, I ran out time to get Mondo working, and had to go
with plan B, which was Kleo Bare Metal Backup:
http://www.kleobackup.net/

This is a GUI wrapper around partimage, similar to Clonezilla Live, but
nicer looking GUI. (Built on an Ubuntu Live CD.)

Kleo, and tools like it, don't really meet my definition of "bare
metal." You have to do a fair bit of work external to the tool to fully
record the partition table, MBR, and all partitions, such that given a
fresh disk, you can faithfully recreate a system.

And being a partition-based tool, there's no file-level granularity.
Temp directories, disposable media files, etc. are all wasting space in
your backup.

It's also difficult to verify the integrity of a backup, short of
restoring the partitions to a new disk and comparing. (partimage
apparently has no built-in verification function.)

These tools also require enough storage space to hold the entire
compressed partition you are backing up. That typically means using a
network share or external disk. Once you've made arrangements to do
that, it's hardly worth the bother to transfer it to removable media.


But the live CD approach does greatly simplify the process, and makes me
wonder if the original design of Mondo using Mindi to create a
customized boot environment is largely unnecessary now that most users
run distribution supplied stock kernels, and many common live CDs work
on most hardware.

When searching around for solutions to my installation problems, I ran
across references to a Mondo live CD that either officially or
unofficially existed at one time, and a few different how-tos for
building a live CD containing Mondo. Clearly there is some interest in
this approach.

  -Tom

------------------------------------------------------------------------------
_______________________________________________
Mondo-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mondo-devel
Loading...