mondo-devel Mailing List for Mondo Rescue

>In theory, up to 510 characters in each parameter (-I '<510 chars>' -E '<510 
>chars>'). You may increase this limit by editing mystuff.h and changing 
>MAX_STR_LEN from 512 to <whatever>.
>
>Perhaps mondo should be able to accept a file specifying paths to be 
>included/excluded.

I think this could be good to put on the list to "might do".

One problem with backing up onto CDs (or DVDs) is they cannot
handle very large disk systems.  So segregating the areas that
are important to backup and the stuff that is expendable - would be
useful.

Lachlan.


-----------------------
Lachlan M. D. Cranswick

Collaborative Computational Project No 14 (CCP14)
    for Single Crystal and Powder Diffraction
  Birkbeck University of London and Daresbury Synchrotron Laboratory 
Postal Address: CCP14 - School of Crystallography,
                Birkbeck College,
                Malet Street, Bloomsbury,
                WC1E 7HX, London,  UK
Tel: (+44) 020 7631 6850   Fax: (+44) 020 7631 6803
E-mail: l.m...@dl...   Room: B091
WWW: http://www.ccp14.ac.uk/

Stan Benoit <sa...@ma...> wrote:-
>I booted from the mindi cd made from mindi.iso, typed compare:
>1. kernel errors flew by to fast from me to read.
>2. mindi dropped back to #

Yikes. Well, I can't reproduce the problem; I've backed up to CD, booted 
from CD; no prob. Backed up to tape, booted from mindi CD; no prob. Would 
you please try today's snapshots & see if you get the same problem as 
before? I'm thinking I've squelched the bug.

-Hugo


_________________________________________________________________
STOP MORE SPAM with the new MSN 8 and get 2 months FREE* 
http://join.msn.com/?page=features/junkmail

On Saturday 30 November 2002 10:20, Kevin Cosgrove uttered:
> Thanks for the generous offer!  I've actually got an 8.2 Mandrake
> PowerPack sitting here in a box.  I was going to install that over my
> company's Christmas break.  But, Mandrake 9.0 came out and now has
> ALSA 0.9 and some other audio stuff I've been itching for.  I've got
> MDK9.0 installed on one computer and it looks great.  I'll be
> installing 9.0 over Christmas break.  If it doesn't work so well,
> then I'll back up to 8.2.

Cool.  MDK 9 is on my test run, seems to be working quite well, especiall=
y=20
with the stable branch.

> P.S.:   Yes, I'm in the US, Oregon actually.

hey, thats just down the road from me.  I'm up in the Seattle area.

--=20
Jesse Keating
For Web Services and Linux Consulting, Visit --> j2Solutions.net
Mondo DevTeam (www.mondorescue.org)

Was I helpful?  Let others know:
 http://svcs.affero.net/rm.php?r=3Djkeating

Thanks for the generous offer!  I've actually got an 8.2 Mandrake 
PowerPack sitting here in a box.  I was going to install that over my 
company's Christmas break.  But, Mandrake 9.0 came out and now has 
ALSA 0.9 and some other audio stuff I've been itching for.  I've got 
MDK9.0 installed on one computer and it looks great.  I'll be 
installing 9.0 over Christmas break.  If it doesn't work so well, 
then I'll back up to 8.2.

P.S.:	Yes, I'm in the US, Oregon actually.

Cheers....

On 29 November 2002 at 23:45, "Hugo Rabson" <hug...@ms...> wrote:

It sounds as if the tests you've run on your hardware have not revealed any 
faults, yet your OS is misbehaving a great deal. You could well be right; 
Mandrake 8.0 could indeed be the problem.

I'm running ArkLinux pre-1.0 and it's more solid than Mandrake has ever 
been. Mandrake 8.0 had a broken kernel and some buggy userland tools, IIRC. 
However, I've enjoyed using Mandrake 8.2; it has behaved very well for me. 
Ditto, Red Hat 7.3.

If you're in the US then I'll gladly mail you some Mandrake 8.2 CD's, if 
that helps.

-Hugo

On Friday 29 November 2002 22:35, Hugo Rabson uttered:
> Having said that, I'd like the patches to be accepted into the mondo 1.=
4x
> and mindi 0.6x soon, if that's okay with Jesse (who's volunteered to
> maintain those branches). I know you've not finished but I like what yo=
u've
> done so far. You've read the code & done sensible things to add the
> features you want.
>
> FYI, 1.60 and 0.80 will be released in mid-December unless a really ugl=
y
> bug suddenly rears its ugly head. If you'd like to submit patches to th=
at
> branch, I'd be very interested in accepting them soon.
>
> Jesse, any comments?

Sounds good to me.  If the patches are to your liking Bruno, I'll patch o=
ur=20
CVS branch, make sure that it doesn't break any other functionality, and =
call=20
it good.  That'll be ample reason enough to roll yet one more rpm/tar rel=
ease=20
set of the stable branch.  I hope to get some testing done on unstable th=
is=20
weekend, between work and playing w/ my new GameCube, so hopefully I can =
give=20
a yea or nea on the work Hugo's been doing.

Bruno, let me know if/when you're happy with the patched functionality. =20
Nothing else is really happening to the stable branches so you've got som=
e=20
time.

--=20
Jesse Keating
For Web Services and Linux Consulting, Visit --> j2Solutions.net
Mondo DevTeam (www.mondorescue.org)

Was I helpful?  Let others know:
 http://svcs.affero.net/rm.php?r=3Djkeating

Bruno Cornec <Bru...@hp...> wrote:-
>Subject: Re: mondo/mindi on ia64 - patch 1 (was: Re: [Mondo-devel] new

The patches look very sane & sensible to me, IMHO. I can't test them on ia64 
hardware (for obvious reasons *g*) so it's very important that I don't 
accept them until you're happy with the way that a patched 
mondo-1.47+mindi-0.67 will run on your ia64 hardware.

Having said that, I'd like the patches to be accepted into the mondo 1.4x 
and mindi 0.6x soon, if that's okay with Jesse (who's volunteered to 
maintain those branches). I know you've not finished but I like what you've 
done so far. You've read the code & done sensible things to add the features 
you want.

FYI, 1.60 and 0.80 will be released in mid-December unless a really ugly bug 
suddenly rears its ugly head. If you'd like to submit patches to that 
branch, I'd be very interested in accepting them soon.

Jesse, any comments?

-Hugo


_________________________________________________________________
The new MSN 8: smart spam protection and 2 months FREE*  
http://join.msn.com/?page=features/junkmail

"Fatal error. Cannot find /dev/md/1's size - is your /etc/raidtab sane?"

You're using Mindi 0.67; try the latest beta, mindi 0.72. If you get the 
same message then check your /etc/raidtab file for bugs.

-Hugo



>From: "Michael E. Labhard" <in...@pa...>
>Reply-To: mon...@li...
>To: mon...@li...
>Subject: [Mondo-devel] Mindi failed.
>Date: Thu, 28 Nov 2002 10:35:18 -0800
>
>Tried the Mindi failsafe kernel, too.  This is on a Gentoo Linux 2.4.19 
>with
>raid 1 devices.  Any help would be appreciated.  Thank you.
>
>-- Michael
><< mindi.err.9850.tgz >>


_________________________________________________________________
MSN 8 with e-mail virus protection service: 2 months FREE* 
http://join.msn.com/?page=features/virus

Lachlan Cranswick <l.m...@dl...> wrote:-
>     Include    /web_disc
>but exclude    /web_disc/ccp14/web_area/web_live/ccp/web-mirrors
>but include    /web_disc/ccp14/web_area/web_live/ccp/web-mirrors/absor

The 'mondo-makefilelist' script isn't that intelligent (yet). I'll stick the 
idea on the mondo-1.7x TODO list. (1.60 is about to come out.)

>Also, how many includes/excludes can Mondo handle?

In theory, up to 510 characters in each parameter (-I '<510 chars>' -E '<510 
chars>'). You may increase this limit by editing mystuff.h and changing 
MAX_STR_LEN from 512 to <whatever>.

Perhaps mondo should be able to accept a file specifying paths to be 
included/excluded.

-Hugo


_________________________________________________________________
MSN 8 with e-mail virus protection service: 2 months FREE* 
http://join.msn.com/?page=features/virus

Kevin Cosgrove <ke...@dO...> wrote:-
>The same hardware was rock solid until I installed
>Mandrake 8.0 about 18 months ago.  Before that it would
>run for months at a time.  Now I have to reboot every
>couple weeks.  I'm hoping life will be better when I
>upgrade to Mandrake 9.0

It sounds as if the tests you've run on your hardware have not revealed any 
faults, yet your OS is misbehaving a great deal. You could well be right; 
Mandrake 8.0 could indeed be the problem.

I'm running ArkLinux pre-1.0 and it's more solid than Mandrake has ever 
been. Mandrake 8.0 had a broken kernel and some buggy userland tools, IIRC. 
However, I've enjoyed using Mandrake 8.2; it has behaved very well for me. 
Ditto, Red Hat 7.3.

If you're in the US then I'll gladly mail you some Mandrake 8.2 CD's, if 
that helps.

-Hugo


_________________________________________________________________
Protect your PC - get McAfee.com VirusScan Online 
http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963

Your logfile says:-
Creating data disk #1...mount: Could not find any loop device, and, 
according to /proc/devices,
       this kernel does not know about the loop device.
       (If so, then recompile or `modprobe loop'.)
shruti: Fatal error. Can't loopmount /tmp/mindilinux/3487/mointpoint.3487\n

Your kernel is broken. It does not support loopmounting. You will need to 
recompile your kernel. Please contact your vendor (Debian) for support. Your 
kernel's requirements are specified in Mondo's manual.

-Hugo


>From: et...@in...
>Reply-To: mon...@li...
>To: mon...@li....
>Subject: [Mondo-devel] (no subject)
>Date: Thu, 28 Nov 2002  16:23 +0530
>
>
>
>Hello,
>
>I'm testing mondo on debian woody. I want to first save the backup on the 
>hard-disk and test the restore before I go request my good (not really) 
>friend to write the CD for me...
>
>Well, something's going wrong and I couldn't find much help from archives 
>of the list or from documentation.
>
>Please help me.
>
>Thanks a ton,
>-shruti
>
>
>
>_____________________________________________________________________________
>If you don't have an account at http://myspace.indbazaar.com ,
>you're missing out on split-second e-mail, 25 MB net space and lots
>of other goodies...
>
>
><< mindi.err.3487.tgz >>
><< mondo-archive.log.gz >>

_________________________________________________________________
Help STOP SPAM with the new MSN 8 and get 2 months FREE*  
http://join.msn.com/?page=features/junkmail

hugo,

I booted from the mindi cd made from mindi.iso, typed compare:
1. kernel errors flew by to fast from me to read.
2. mindi dropped back to #

Review log files:
Attempted to mount /dev/hdax, so i could copy the log files to send you. 
No go. I did veiw them, the log file was short and looked like dmesg 
truncated. 

Plan: 
Burn another boot cd see if maybe the last burn was bad. And figure out 
how to get these logs to you. No floppy in this box. :(

Backup went without issue.


-- 
Stan Benoit

On Thursday 28 November 2002 22:27, Reckhard, Tobias uttered:
> I wouldn't expect that to be easy with stock NFS implementations, becau=
se:
> 1. NFS and portmapper use UDP, while AFAIK SSH can only tunnel TCP.
> 2. NFS, as an RPC service, isn't guaranteed to use one and the same por=
t,
> but you have to specify the port to be forwarded in SSH. However, most =
NFS
> implementations default to port 2049, so this isn't as much of a proble=
m as
> #1.

Thats old news.  Current NFS implimentations use TCP, and with the use of=
 tcp=20
wrappers and portmapper, you can force NFS to use the same tcp port for=20
everything.  I've been told that it's not that difficult to do, if it doe=
sn't=20
do it already.

--=20
Jesse Keating
For Web Services and Linux Consulting, Visit --> j2Solutions.net
Mondo DevTeam (www.mondorescue.org)

Was I helpful?  Let others know:
 http://svcs.affero.net/rm.php?r=3Djkeating

> > If you're interested, we could go over the steps to tunnel 
> NFS through SSH.  
> > It sounds like a fun project (:
> 
> Thanks for the offer. I'll think about it.

I wouldn't expect that to be easy with stock NFS implementations, because:
1. NFS and portmapper use UDP, while AFAIK SSH can only tunnel TCP.
2. NFS, as an RPC service, isn't guaranteed to use one and the same port,
but you have to specify the port to be forwarded in SSH. However, most NFS
implementations default to port 2049, so this isn't as much of a problem as
#1.

I'd expect a CIFS mount (i.e. using Samba) to be easier to do. But I don't
know if mondo supports it.

As for securing NFS, you could use IPSec to do (at least part of) that.
Depends on what you're afraid of, of course. And I prefer to avoid NFS
wherever possible, too.

Cheers,
Tobias

Tried the Mindi failsafe kernel, too.  This is on a Gentoo Linux 2.4.19 w=
ith=20
raid 1 devices.  Any help would be appreciated.  Thank you.

-- Michael

Le mer 27/11/2002 =C3=A0 18:23, Bruno Cornec a =C3=A9crit :

I am one of the new geek on automake, autoconf ... I could see If we
could do some tests in ./configure to make it more IA64 compliant :)

Well new beta (1.4x) mondo already do a bunch of tests ...
May be you could try those one.
--=20
         ,,,
   ____()() }                                          .--.=20
   =3D=3D=3D=3D=3D=3D=3D (                     Duckx                |o_o |
           \\     \               ***                 |:_/ |
           ) \____=3D=3D=3D.             *                 //   \ \
          /    (  )\(           Oinc !!             (|     | )
         //     \ [%]                               /'\_   _/`\        =20
         ^^^    ^^^          @@@@@@@@@@@@@@@        \___)=3D(___/
         \_/----\_/     =20
                      mail : duckx at mezimail.com

Hi,

> > This is the 2.4.18 kernel and, judging from the kernel
> > description, all known bugs of nfs should have been fixed. Please feel
> > asured that this is not the case )-: as we can reproduce trashing of the
> > file contents as soon as you push more that one gig.
> >
> > Hence: Hands off concerning nfs and big files....
> 
> Hrm, so this is outside of mondo's control...  I wonder if any bugfiles have 
> been reported.  Can you let me know the steps you took to reproduce this?  
> I'd like to bring it up to the Red Hat kernel developers I know.

I'm about to check this out. It seems moving a big file into an nfs-mounted
directory is enough to corrupt it past the Gig barrier. The other alternative
I'm also following right now is that the ext-filesystem on the mentioned
computer is screwed. We'll see....

So long,
	Thomas

On Thursday 28 November 2002 08:50, Thomas Richter uttered:
> This is the 2.4.18 kernel and, judging from the kernel
> description, all known bugs of nfs should have been fixed. Please feel
> asured that this is not the case )-: as we can reproduce trashing of th=
e
> file contents as soon as you push more that one gig.
>
> Hence: Hands off concerning nfs and big files....

Hrm, so this is outside of mondo's control...  I wonder if any bugfiles h=
ave=20
been reported.  Can you let me know the steps you took to reproduce this?=
 =20
I'd like to bring it up to the Red Hat kernel developers I know.

--=20
Jesse Keating
For Web Services and Linux Consulting, Visit --> j2Solutions.net
Mondo DevTeam (www.mondorescue.org)

Was I helpful?  Let others know:
 http://svcs.affero.net/rm.php?r=3Djkeating

On Thu, 2002-11-28 at 11:46, Jesse Keating wrote:

> Well, monitas is somewhat at a standstill I think.  Mondo got really popular 
> as of late, so a lot of our (the dev team) attention was turned toward mondo.
> 
> Work should continue on monitas, but I think it needs to be more of a group 
> effort.  (and we should wait for mondo to settle down again)
> 

Ok. I'll keep an eye out for announcements.


> As to your needs.  Have you thought about tunneling NFS through SSH?  This 
> will give you a nice secure method of backing up the files, and you can even 
> use this NFS mount for your temp and scratch directories, so that it won't 
> take up very much space on your system.

No, I haven't thought about it, mainly because I'm not too fond of NFS.
I'm not too worried about the security aspects, I just don't like NFS. I
suppose that if that's all that available right now, then that's what
I'll have to use though. That or get a CD burner for that machine too,
which may not be a bad idea anyway.

> 
> If you're interested, we could go over the steps to tunnel NFS through SSH.  
> It sounds like a fun project (:

Thanks for the offer. I'll think about it.


-- 
Dion Vansevenant
Smaller Solutions
di...@sm...
519-872-0934

Hi folks,

I reported about a week ago problems concering mondo and backup over nfs,
and I reverted the backup strategy to mindi directly for this purpose,
see [mondo DVD trouble].
While doing so, we found a problem in nfs - unfortunately. It seems that
nfs has problems to write huge files over the net, namely it seems to have
a 1GB problem. This is the 2.4.18 kernel and, judging from the kernel
description, all known bugs of nfs should have been fixed. Please feel asured
that this is not the case )-: as we can reproduce trashing of the file
contents as soon as you push more that one gig.

Hence: Hands off concerning nfs and big files.... 

I haven't tried mondo again since we really don't have enough hard disk
space on the machine *locally* that we want to backup /-: but at least
the trouble maker seems to be identified for the moment.

Greetings,
	Thomas

On Thursday 28 November 2002 08:04, Dion Vansevenant uttered:
> My firewall/gateway/web server/database server/mail server machine just
> doesn't have enough disk space to backup to itself. I don't like using
> NFS, so I was hoping to pipe the stream through an ssh connection. The
> client/server Mondo sounds like a better solution.
>
> Can I still do what I want in the meantime? Anyone else doing something
> like this?

Well, monitas is somewhat at a standstill I think.  Mondo got really popu=
lar=20
as of late, so a lot of our (the dev team) attention was turned toward mo=
ndo.

Work should continue on monitas, but I think it needs to be more of a gro=
up=20
effort.  (and we should wait for mondo to settle down again)

As to your needs.  Have you thought about tunneling NFS through SSH?  Thi=
s=20
will give you a nice secure method of backing up the files, and you can e=
ven=20
use this NFS mount for your temp and scratch directories, so that it won'=
t=20
take up very much space on your system.

If you're interested, we could go over the steps to tunnel NFS through SS=
H. =20
It sounds like a fun project (:

--=20
Jesse Keating
For Web Services and Linux Consulting, Visit --> j2Solutions.net
Mondo DevTeam (www.mondorescue.org)

Was I helpful?  Let others know:
 http://svcs.affero.net/rm.php?r=3Djkeating

On Wednesday 27 November 2002 10:14, Argggggggghhhhhhhhh uttered:
> Real Name :Fr=C3=A9d=C3=A9ric Brin .... Aka a good french programming F=
rog :)

yes! thats it!

--=20
Jesse Keating
For Web Services and Linux Consulting, Visit --> j2Solutions.net
Mondo DevTeam (www.mondorescue.org)

Was I helpful?  Let others know:
 http://svcs.affero.net/rm.php?r=3Djkeating

I was perusing the archives to see if I would send my mondo backups
through an ssh pipe when I stumbled across a proposal for a
client/server version of Mondo back April. Has anything been done with
this? Sounded really interesting.

My firewall/gateway/web server/database server/mail server machine just
doesn't have enough disk space to backup to itself. I don't like using
NFS, so I was hoping to pipe the stream through an ssh connection. The
client/server Mondo sounds like a better solution.

Can I still do what I want in the meantime? Anyone else doing something
like this?

Thanks.

-- 
Dion Vansevenant
Smaller Solutions
di...@sm...
519-872-0934

Hugo,

The backup went well, no problems. Please review attachment. 

I'll try a nuke restore see how it goes. 

-- 
Stan Benoit

     
               
Hello, 

I'm testing mondo on debian woody. I want to first save the backup on the hard-disk and test the restore before I go request my good (not really) friend to write the CD for me... 

Well, something's going wrong and I couldn't find much help from archives of the list or from documentation.

Please help me. 

Thanks a ton,
-shruti
               


_____________________________________________________________________________
If you don't have an account at http://myspace.indbazaar.com , 
you're missing out on split-second e-mail, 25 MB net space and lots
of other goodies...

hr> Your SCSI bus resets itself whenever it feels like it.

	Actually, this is the first SCSI reset (set of them) that
	I've seen in years.  Back then I was using an external
	QIC-150 drive on the system that seemed a bit flakey.

hr> This is very odd. It's almost as if there's something not quite
hr> right with your PC.

	I've been looking.  My next thought is to pull half of my
	RAM and recompile a kernel again looking for RAM problems.

	The same hardware was rock solid until I installed
	Mandrake 8.0 about 18 months ago.  Before that it would
	run for months at a time.  Now I have to reboot every
	couple weeks.  I'm hoping life will be better when I
	upgrade to Mandrake 9.0

hr> As you'll know from having read Mondo's source code, Mondo does
hr> no clever system calls. It uses tools such as afio, mkisofs and
hr> cdrecord to do the bulk of the work. If cdrecord has a bug big
hr> enough to cause a SCSI reset then it would probably have been
hr> experienced by more of Mondo's users than just 1.

	I just got done buring 9-10 CDRs w/o any troubles.  All
	the CDR checksums match the ISO image checksums I made.

hr> Would you care to investigate?

	Yup.  I've been doing some of that, e.g. kernel compiles,
	RAM/disk thrashing.  Were you after something specific
	that you think I should do?

	Cheers....