mondo-devel Mailing List for Mondo Rescue

Hey Bruno,

Thank you!  It never occurred to me that whitespace might be significant.

So mondo ran to completion.  Took about 4 hours and generated 23 
DVD-sized files.  All pretty reasonable.

I'm happy to have mondo working again!

Howard

On 5/10/2025 6:43, Bruno Cornec via Mondo-devel wrote:
> Howard Spindel said on Fri, May 09, 2025 at 06:56:13PM -0700:
>
>> mount -t tmpfs -o size=5g tmpfs /mnt/ramdisk
>> mondoarchive -Oi -d /mnt/mondo -s 4700m -9 -N -T "/mnt/ramdisk/" -E 
>> "/tmp | /home/rear-tmp | /home/mondo-tmp | /home/timeshift"
>
> That exclude line is wrong. Please use:
> -E "/tmp|/home/rear-tmp|/home/mondo-tmp|/home/timeshift"
>
> The space is taken as a part of the directory name in the way you used.
>
>> DBG1: [Main] libmondo-cli.c->process_switches#591: WARNING ! 
>> /home/mondo-tmp  doesn't exist
>
> See the extra space ?
>
> Best regards,
> bruno.

Howard Spindel said on Fri, May 09, 2025 at 06:56:13PM -0700:

>mount -t tmpfs -o size=5g tmpfs /mnt/ramdisk
>mondoarchive -Oi -d /mnt/mondo -s 4700m -9 -N -T "/mnt/ramdisk/" -E 
>"/tmp | /home/rear-tmp | /home/mondo-tmp | /home/timeshift"

That exclude line is wrong. Please use:
-E "/tmp|/home/rear-tmp|/home/mondo-tmp|/home/timeshift"

The space is taken as a part of the directory name in the way you used.

>DBG1: [Main] libmondo-cli.c->process_switches#591: WARNING ! 
>/home/mondo-tmp  doesn't exist

See the extra space ?

Best regards,
bruno.
-- 
Des infos sur la musique ancienne  -- http://www.musique-ancienne.org
Des infos sur les logiciels libres -- http://www.HyPer-Linux.org
Home, sweet musical Home -- Lover of Andromède, Béatrice, Early Music, 
Josquin, Linux, Mélisande, Recorder, and Ségolène (not in that order)

Hello Bruno,

On 5/9/2025 7:11, Bruno Cornec via Mondo-devel wrote:
>
>> I tried ever-increasing sizes for EXTRA_SPACE.  It did not work until 
>> mindi.conf has the following:
>>
>> BOOT_SIZE=65536
>
> I don't have that parm in the code, so I think it's not creating any
> impatc.

I found it on the web somewhere.  It doesn't seem to hurt anything to 
have it in there.
>
>
>> EXTRA_SPACE=1000000
>
> That's quite a lot.
>> It might have worked with a smaller size, but I didn't try anything 
>> between 500000 and 10000000.
>
> I think that 600MB are generally enough. Anyway, that doesn't hurt too
> much anyway.
>
> However, I remarked that there is an underlying bug with the creation
> that I should work on to avoid the need to raise that number that much.
> Ultimate goal would be to compute it, without you needing to change it.
>
> Bruno.

Computing it would be very cool.

I have been running my first mondo backup, and while it seems to be 
progressing I fear something is wrong.  It has been running for about 17 
hours and there are now 126  4.6GB (DVD-sized) files created.  It's 
still going.  That's a total of 567GB.

For comparison ReaR runs in 2-3 hours and generates a tar file of 75GB 
(on the same machine with the same exclusions).

Here is the shell script I use to launch mondo:

/home/howard/dd/mount-mondo
mount -t tmpfs -o size=5g tmpfs /mnt/ramdisk
mondoarchive -Oi -d /mnt/mondo -s 4700m -9 -N -T "/mnt/ramdisk/" -E 
"/tmp | /home/rear-tmp | /home/mondo-tmp | /home/timeshift"
umount /mnt/ramdisk
umount /mnt/mondo

-------------------

Looking in mondoarchive.log I see it appears to be backing up the 
directory /home/timeshift, which is huge and could easily account for 
what I'm seeing.  But you can see in the invocation line (and in 
mondoarchive.log) that /home/timeshift is excluded. ReaR has 
/home/timeshift excluded.

Here are  relevant lines from mondoarchive.log:
DBG1: [Main] libmondo-devices.c->mr_make_devlist_from_pathlist#2114: 
exclude_paths is now '|/tmp | /home/rear-tmp | /home/mondo-tmp | 
/home/timeshift'

DBG4: [Main] libmondo-files.c->size_of_all_biggiefiles_K#803: 
/home/timeshift/snapshots/2023-07-27_23-00-35/localhost/boot/initramfs-0-rescue-5b4392de242040cf88adc3bd3df13453.img 
--> 121163 K
DBG4: [Main] libmondo-files.c->size_of_all_biggiefiles_K#805: After 
adding 
/home/timeshift/snapshots/2023-07-27_23-00-35/localhost/boot/initramfs-0-rescue-5b4392de242040cf88adc3bd3df13453.img, 
scratchL+121163 now equals 123411725
DBG4: [Main] libmondo-files.c->size_of_all_biggiefiles_K#803: 
/home/timeshift/snapshots/2023-07-27_23-00-35/localhost/boot/initramfs-5.14.0-162.6.1.el9_1.x86_64.img 
--> 73234 K
DBG4: [Main] libmondo-files.c->size_of_all_biggiefiles_K#805: After 
adding 
/home/timeshift/snapshots/2023-07-27_23-00-35/localhost/boot/initramfs-5.14.0-162.6.1.el9_1.x86_64.img, 
scratchL+73234 now equals 123484959

I also found this:

DBG1: [Main] libmondo-cli.c->process_switches#591: WARNING ! /tmp  
doesn't exist
DBG1: [Main] libmondo-cli.c->process_switches#591: WARNING ! 
/home/rear-tmp  doesn't exist
DBG1: [Main] libmondo-cli.c->process_switches#591: WARNING ! 
/home/mondo-tmp  doesn't exist
DBG1: [Main] libmondo-cli.c->process_switches#596: WARNING ! 
/home/timeshift doesn't exist

which is weird as those directories definitely exist:

[root@server2 /]# ls -l home
total 16
drwx------   3 bw        bw         163 Jul 27  2023 bw
drwx------   3 cc        cc         163 Aug 28  2024 cc
drwx------   3 christina christina  163 Aug 28  2024 christina
drwx------   3 debra     debra      163 Jul 27  2023 debra
drwx------. 55 howard    howard    8192 May  9 17:47 howard
drwxr-xr-x   3 root      root        38 May  8 21:21 mondo-tmp
drwxr-xr-x   8 mysql     mysql     4096 May  7 19:29 mysql
drwxr-xr-x   2 root      root         6 May  8 17:06 rear-tmp
drwxr-xr-x   2 serviio   serviio      6 Nov 17  2023 serviio
drwxr-xr-x   2 root      root         6 May  8 01:25 snapstore
drwxr-xr-x   9 root      root       199 May  9 18:04 timeshift

I tried to attach mondoarchive.log, but it is too big to get through 
(and zip files don't work).

Am I doing something wrong syntactically in failing to get 
/home/timeshift excluded?  I did try modifying the invocation line to 
add a / at the end of each excluded directory, but it made no 
difference. I also tried excluding /home/timeshift/* with no joy.

Thank you for all your help!
Howard

Hello Howard,

Howard Spindel said on Thu, May 08, 2025 at 04:46:33PM -0700:

>I tried ever-increasing sizes for EXTRA_SPACE.  It did not work until 
>mindi.conf has the following:
>
>BOOT_SIZE=65536

I don't have that parm in the code, so I think it's not creating any
impatc.

>EXTRA_SPACE=1000000

That's quite a lot. 

>It might have worked with a smaller size, but I didn't try anything 
>between 500000 and 10000000.

I think that 600MB are generally enough. Anyway, that doesn't hurt too
much anyway.

However, I remarked that there is an underlying bug with the creation
that I should work on to avoid the need to raise that number that much.
Ultimate goal would be to compute it, without you needing to change it.

Bruno.
-- 
Des infos sur la musique ancienne  -- http://www.musique-ancienne.org
Des infos sur les logiciels libres -- http://www.HyPer-Linux.org
Home, sweet musical Home -- Lover of Andromède, Béatrice, Early Music, 
Josquin, Linux, Mélisande, Recorder, and Ségolène (not in that order)

Thank you for the response, Bruno.

I tried ever-increasing sizes for EXTRA_SPACE.  It did not work until 
mindi.conf has the following:

BOOT_SIZE=65536
EXTRA_SPACE=1000000

It might have worked with a smaller size, but I didn't try anything 
between 500000 and 10000000.

The machine is a Virtual Box Rocky Linux install, with 16GB main memory 
assigned.

Does this seem reasonable?

Thank you,
Howard




On 5/8/2025 7:57, Bruno Cornec via Mondo-devel wrote:
> Howard Spindel said on Sat, May 03, 2025 at 06:32:21PM -0700:
>> Log files attached.
>
> Thanks.
>
>> Please note that I tried to send a zip file as an attachment, but it 
>> bounced with the following:
>> <mon...@li...>: host 
>> mx.sourceforge.net[216.105.38.6]
>>    said: 550 Blacklisted file extension detected (in reply to end of 
>> DATA
>>    command)
>
> Probably a new policy from sourceforge.net. I'm considering moving this
> ML to my own ML server which I already use for other communities.
>
>> cp: error writing 
>> '/tmp//mondo.tmp.zNZzns/mountpoint.65782//usr/lib/firmware/mellanox/mlxsw_spectrum-13.2008.2438.mfa2.xz': 
>> No space left on device
>
> Seems the boot image is full.
> Can you try inceasing the size with EXTRA_SPACE please:
> http://trac.mondorescue.org/wiki/FAQ#Q50MondoRescuecomplainsaboutfilesystemfullonRHEL7 
>
>
> You can go up to the size of a media if you want, but 500MB should be
> sufficient.
> cd
> Best regards,
> Bruno.

Howard Spindel said on Sat, May 03, 2025 at 06:32:21PM -0700:
>Log files attached.

Thanks.

>Please note that I tried to send a zip file as an attachment, but it 
>bounced with the following:
><mon...@li...>: host mx.sourceforge.net[216.105.38.6]
>    said: 550 Blacklisted file extension detected (in reply to end of DATA
>    command)

Probably a new policy from sourceforge.net. I'm considering moving this
ML to my own ML server which I already use for other communities.

>cp: error writing '/tmp//mondo.tmp.zNZzns/mountpoint.65782//usr/lib/firmware/mellanox/mlxsw_spectrum-13.2008.2438.mfa2.xz': No space left on device

Seems the boot image is full.
Can you try inceasing the size with EXTRA_SPACE please:
http://trac.mondorescue.org/wiki/FAQ#Q50MondoRescuecomplainsaboutfilesystemfullonRHEL7

You can go up to the size of a media if you want, but 500MB should be
sufficient.

Best regards,
Bruno.
-- 
Des infos sur la musique ancienne  -- http://www.musique-ancienne.org
Des infos sur les logiciels libres -- http://www.HyPer-Linux.org
Home, sweet musical Home -- Lover of Andromède, Béatrice, Early Music, 
Josquin, Linux, Mélisande, Recorder, and Ségolène (not in that order)

Hello,

Carl Wignall via Mondo-devel said on Wed, May 07, 2025 at 11:56:06AM +0100:
>I have been using mondo for many years now, mainly with Centos6, more 
>latterly Rocky 8.x and it has been previously extremely useful and 
>reliable.

Good news ;-)

>Lately I experienced restore errors on any (Rocky) version > 8.5 and 
>was today trying to replicate that as i saw recent updates to the 
>Rocky 8 folder on your ftp.

Well less good news :-(

>     Problem: conflicting requests
>      - nothing provides buffer needed by mondo-3.2.2-1.rocky8.x86_64 
>from mondorescue

That should now be solved, as I have provided buffer for rocky 8.

>Previously i have been using the Centos8 repo, which worked upto and 
>including Rocky version 8.5.  Higher versions of Rocky the backup 
>appears to work, the restore appears to work, but the target fails to 
>boot
>    Timed out waiting for device /dev/mapper/rl-home
>    Dependency failed for /home
>    Dependency failed for Local File Systems
>    Dependency failed for Mark the need to relable after reboot
>It then tries to drop you into Emergency mode, but that fails to 
>complete also!
>I was hoping today to extract the mindi log for you but was scuppered 
>with the above issue

So, I also updated the repo with the latest version of mondo that
contains additional fixes wrt what you have used (I think) up to now.
(now you seem to be using that same version on rocky 9)

>Sorry for the two problems in one email, but i'm hoping we can get to 
>the bottom of both as i would really love to get Mondo working again 
>on Rocky8.   Incidentally i tried a test build/backup&restore using 
>Rocky 9 and that also fails with the above failure to mount errors.

Well, as I often underline, without log files, I've difficulties
understanding what could go wrong.

Maybe something changed in the way FS are managed starting with rocky >
8.4. Pleas provide me wit the archive logs, and also the restore logs if
you can get them (see http://www.mondorescue.org/support.shtml)

With more details, I hope to be able to fix the issues you've
encountered.

Greetings,
Bruno.
-- 
Des infos sur la musique ancienne  -- http://www.musique-ancienne.org
Des infos sur les logiciels libres -- http://www.HyPer-Linux.org
Home, sweet musical Home -- Lover of Andromède, Béatrice, Early Music, 
Josquin, Linux, Mélisande, Recorder, and Ségolène (not in that order)

Hi,

I have been using mondo for many years now, mainly with Centos6, more 
latterly Rocky 8.x and it has been previously extremely useful and reliable.
Lately I experienced restore errors on any (Rocky) version > 8.5 and was 
today trying to replicate that as i saw recent updates to the Rocky 8 
folder on your ftp.

Unfortunately i still get: (using the Rocky8 repo)
     [root@******* *****]# yum install mondo
     Last metadata expiration check: 0:00:47 ago on Wed 07 May 2025 
11:16:55 BST.
     Error:
      Problem: conflicting requests
       - nothing provides buffer needed by mondo-3.2.2-1.rocky8.x86_64 
from mondorescue
     (try to add '--skip-broken' to skip uninstallable packages or 
'--nobest' to use not only best candidate packages)


Previously i have been using the Centos8 repo, which worked upto and 
including Rocky version 8.5.  Higher versions of Rocky the backup 
appears to work, the restore appears to work, but the target fails to boot
     Timed out waiting for device /dev/mapper/rl-home
     Dependency failed for /home
     Dependency failed for Local File Systems
     Dependency failed for Mark the need to relable after reboot
It then tries to drop you into Emergency mode, but that fails to 
complete also!
I was hoping today to extract the mindi log for you but was scuppered 
with the above issue

Sorry for the two problems in one email, but i'm hoping we can get to 
the bottom of both as i would really love to get Mondo working again on 
Rocky8.   Incidentally i tried a test build/backup&restore using Rocky 9 
and that also fails with the above failure to mount errors.

Cheers,
Carl

mindi.log attached.

Howard

On 5/2/2025 1:57, Bruno Cornec via Mondo-devel wrote:
> Hello,
>
> Howard Spindel said on Thu, May 01, 2025 at 12:56:36AM -0700:
>> I'd really appreciate a pointer on how to fix this error.
>
> Well, I'd appreciate a log file in order to help you more :-)
> http://www.mondorescue.org/support.shtml
>
> Bruno.

Thank you for the pointer, Bruno.

Log files attached.

Please note that I tried to send a zip file as an attachment, but it 
bounced with the following:

<mon...@li...>: host mx.sourceforge.net[216.105.38.6]
     said: 550 Blacklisted file extension detected (in reply to end of DATA
     command)

I am therefore going to follow this up with another email with mindi.log attached.


Thank you,
Howard

On 5/2/2025 1:57, Bruno Cornec via Mondo-devel wrote:
> Hello,
>
> Howard Spindel said on Thu, May 01, 2025 at 12:56:36AM -0700:
>> I'd really appreciate a pointer on how to fix this error.
>
> Well, I'd appreciate a log file in order to help you more :-)
> http://www.mondorescue.org/support.shtml
>
> Bruno.

Hello,

Howard Spindel said on Thu, May 01, 2025 at 12:56:36AM -0700:
>I'd really appreciate a pointer on how to fix this error.

Well, I'd appreciate a log file in order to help you more :-)
http://www.mondorescue.org/support.shtml

Bruno.
-- 
Des infos sur la musique ancienne  -- http://www.musique-ancienne.org
Des infos sur les logiciels libres -- http://www.HyPer-Linux.org
Home, sweet musical Home -- Lover of Andromède, Béatrice, Early Music, 
Josquin, Linux, Mélisande, Recorder, and Ségolène (not in that order)

After not using mondo for several years, I've wanted to get it working 
again.

On Rocky Linux 9.5, I issue the following command (running as root):

mondoarchive -Oi -d /tmp/backup/ -s 4g -T "/tmp" -E "/tmp"

mondo runs for quite a while (runs mindi)  but aborts with the message:

rm: cannot remove '/tmp/mondo.tmp.NKhCxR/mountpoint.165168': Device or 
resource busy
Execution run ended; result=254

Note that if I manually try to delete that directory still Linux tells 
me Device or resource busy.

lsof doesn't tell me anything about a program that has the file open.
mount shows the above directory mounted, but with (deleted) after it.  I 
cannot umount.

If I reboot the machine, the directories under /tmp/mondo* are gone.

--------------------------------

If I try to run mondo from the interactive TUI, it bails immediately 
with result=-1.  Failed to create global scratch directory /mnt/xfer.  I 
have no idea why mondo is choosing /mnt/xfer instead of something like 
/tmp.  Possibly the reason mondo complains that it can't create 
/mnt/xfer is that /mnt/xfer is actually an existing directory.

I'd really appreciate a pointer on how to fix this error.

Thank you,
Howard