I am not very good with Linux/Ubuntu (I'm a newbie) but I will explain my problem as best I can and I hope someone can help me.
I have an Ubuntu Server, version 6.10 and installed on it is a Postgresql database which I have about 100 remote clients constantly accessing it (sending and requesting data). I also have the latest version of Mondo installed and I recently made a live back up (when I did the backup to .iso images, I left Postgresql running and the box connected to the database, with the remote clients still using the database).
Mondo made 6 .iso images which I later moved to a Windows XP Pro machine and burned CD's with them (there were no errors when buring the iso images to CDs).
I then took an exact duplicate machine (both are Dell's with same exact hardware inside) and booted off the 1st CD and began to 'restore' the backup on this new machine using the Mondo/Mindi GUI.
During the process after the 1st CD just started to be 'restored' to this new machine, I got an error that mentioned 'out of memory' 'Kill process 12420 (mondorestore) score 164 and children' ... and left me at a 'BusyBox 1.00' prompt....
My question is; Did this 'restore error' happen because I made the mondo backup with postgresql live and running while clients where still accessing the machine?
Do you think if I shut down Postgresql and pulled the machine off the internet, and then did a Mondo back up, the 'restore' would then work?
It is entirely possible that this occurred. May I also ask one question? Why did you choose Ubuntu, as well as version 6.10? Besides being nearly 2 years old, I never found Ubuntu to be the best setup for a server. Others may disagree. I would recommend SUSE or CentOS (Red Hat).
I had an outside programmer set up the server/database for me; and it was about 2 years ago when it was done. It's always ran good and has done what I needed it to, that is why I haven't updated it. 'Dont fix whats not broken' is what I always say when it comes to computers/software...
I still work with this outside programmer, but I want to depend on them less and less now; that is why I want to back the server up incase I mess it up...
The out of memory error looks more like a problem with the restore application than the backup. If looks lik it was unable to find sufficient memory to perform the tasks is trying to run, or that it could not fork a child process becasue there was no memory available. How much RAM is in the machine you are restoring to?
I tried again to back up to ISOs and this time I stopped the server and did it.
But again when I go to 'restore' it to new machine (which mirrors the original machine), I still get the same errors ....
Does anyone have any ideas?
'reborg', you asked how much memory the machine has that I am restoring to, and it has 1.0 Gig ; which is the same amount of RAM the original server has ....
Is there a reason why I would need more ram?
10.8.5. Q: When I restore after booting from the media, I sometimes get errors like, "Running out of memory" or "Segmentation fault". What is going on?
From their FAQ:
A: It sounds as if you are running out of disk space, probably ram disk space. Type 'df -m' to see which partitions are running low on space. Please send as much information as you can to the mondorescue mailing list. This problem is believed to have been fixed in 1.63 and 1.71.
If you have the ability to do so, you can take a "poor man's" backup of the machine to clone it by putting the other disk into an enclosure (USB or Firewire) and then attaching it to the original machine.
If you can take downtime, drop to single user mode and copy the physical device to the other physical device such as:
dd if=/dev/sda of=/dev/sdb
where "if" is the input file (the disk you are copying) and "of" is the output file, the place you are copying it to.
This will copy EVERYTHING from the first disk to the second disk, barring failures or hardware issues. However, it will copy blank space and will take a long time. Figure on about 6hrs easily.
Also, if you accidentally screw up the if and of parts, you will have destroyed your machine (No joke, all data will be lost)