Agree, tho I do not verify my Clonezilla backups. Maybe that’s one reason I try to always have a grandfather, father, and son backup. I am hoping with Timeshift, one of the 4 will be able to restore the OS.
Hope is not a strategy, even if Timeshift works like a charm.
Do you mean a backup of the backup? This is useless if the initial backup is faulty or incomplete. If you want to be certain, there is, IMHO, no way around verification.
Lets get it clear… do you mean verify the software works (once only), or verify every backup run?
Yes, only once. As proof of process. You might have omitted vital files accidentally, or the restore method might fail for any unexpected reason.
Of course, there are more things to consider.
- Use more than one backup medium, e.g. one for the odd days and the other for even days.
- Keep the backup medium apart, never close to the original medium.
- …
First, Ernie, best of luck. Looks like you are doing what you should be doing, weight is important. Us oldies (91), have to stick together.
I am using Pi’s for more things lately, now that nvme drives are supported. I have a Pi5 running OMV (OpenMediaVault). It has 2ea 1TB drives in an mdadm mirror. Pironman5 Max. Yup, about 375. I boot from another nvme on USB3 (more , but I already had it). I think this is still cheaper than most off the shelf NAS boxes and I have full control. Supports smb, nfs, rsync, smbmount for d&d, etc. I used to use a large Dell workstation running TrueNAS but we are going to move to an apartment so I have to downsize.
For full system backups, I find that Veeam agent for Linux (community) is extremely powerful. You can do live drive/partition/file backups. You can schedule it and use an option to retain x number of images. Requires creating a user account. It has a rather basic UI or you can use cli. We keep a monthly, full image, off site. For local machine, I do monthly Timeshift backups. I have 3 linux systems active (triple boot?).
We also use curl to a virtual debian system running nextcloud, this for nightly data backups.
Testing backups.: Statistical sample sizes say that if you want to be 100% sure that everything in the backup is perfect, you have to test accordingly (100%). Most of the time, I want to be “pretty sure” that the backup isn’t corrupt, so I mount the backup (Veeam) and look at a couple of files. Every couple of months, I put a spare drive in the system, boot from the recovery stick and do a full drive restore. So far, it has never failed. Murphy says that if you don’t test, they will all fail when you need them. I currently run 4 Pi 5’s. Some habits are expensive. One runs PiHole 24/7. Another runs nginx 24/7, hosting 3 web sites, it has a public IP address. Third runs OMV and the 4th is just for fun, like Plex. Most are birthday presents, if I don’t specify something, I get shirts ![]()
I have not contributed to this discussion before now as I would never use command line to do a backup when the tools exist in the gui. Whybthe extra pain
But just read an item about the new version of linux mint 22.3 and some of the new features.
Especially backup.
“There have also been improvements in Mint’s backup programs, Timeshift and Mint Backup. TimeShift takes snapshots of system files (root filesystem, critical config), allowing the operating system to be rolled back to a previous working state after a bad update, driver, or config change.”
And
“Timeshift, Mint’s backup program, can also now pause and resume snapshots. This capability makes scheduled backups easier if something goes wrong on the server or cloud you’re using to back up your system configuration. Mint Backup, with its new ‘Include All’ button, makes it simple to add all your hidden files and folders to your backup”
Details can be found at
No not at all. 3 different backups taken on 3 different dates.
And of course, tested procedure and actual use of restoring a backup a few times.
Not testing a restore procedure is a way of looking for trouble.
This is why I asked. Then they are more siblings than grandparent, parent, and child. If your backup process is faulty, then all of them are as well.
Well said!
![]()
Yes. We have off/on site kits. We could buy new hardware and restore the OS and 4 virtual servers, anywhere. Tested and documented. I also keep a mirror of that system in my SOHO, would run but be slow. Untested backup is similar to no backup. Timeshift is handy, I recovered a (self made) non booting Debian system with it.
Just read the article myself. A couple comments good and bad stood out to me.
“Mint remains boring but good” and “provides a conservative, well‑tested foundation with current hardware support” and last
“This update makes Mint, once more, one of the best desktops for people who just want their PC to work.”
I have 2 desktops. so I might perform a fresh install of Mint 22.3 on one and try the upgrade path on the other desktop. I really don’t need any of the new features, but enjoy staying current. Mint 22 is supported until 2029.
Before making any major changes, I will use Clonezilla to make a backup of the current
system. A backup procedure that is well tested and used several times.
I suppose the older a utility the better it has been tested.
One of the oldest would be the original Unix dump and restore commands.
Does anyone use dump and restore?
Had not hear of that procedure. But tar must be pretty old too.
Seems like I remember using the tar command to dump and restore a database back in the
late 1990’s.
Tar and dump/restore were both there in the first Unix I used … BSD4.2 in about 1980. Dump has automated incremental dumps. We used to dump overnight to a tape… incrementals for 5 days then a full dump on weekend. To recover you had to put in the last full dump , then all the incrementals since then in order.
Have a look at man dump … it will be there in your Linux.
Can someone restore our main road by dumping some tar in these holes
![]()
Just playing with the words. Who thought of them.
Yes, same here. We tar the database to tape, but I believe the Unix Admin decided to
tar the whole file every backup. I was wondering what tar stool for. Tape ARchive.
Sorry to hear about your health issues @ernie …
I’ve mentioned on some other thread - 6 weeks ago my wife went to the ER and was diagnosed as having a small heart attack, and Type 2 diabetes… only 61…
Kinda bizarre - my family has a history of dodgy tickers and thought I’d cop it before her (I’m 2 years older)…
As for salt - I’ve hardly ever added it to food myself - but - I nearly always add black pepper to food… I have high blood pressure and cholestorol and I’m on 3 meds for that… I’ve noticed something recently - like when I donate blood - if I drive the car there - I end up with high blood pressure - if I ride my Harley to the blood bank - it’s low / healthy… Lesson: ride motorcycles more often!
My mum was a bit of a health food nut - she got it from her Dad… First time I ate with my missus’ family I was a bit shocked, they’d drown their food in salt… We never had salt on the table - sometimes my mum would put kelp tablets for us to swallow with diluted apple cider vinegar…
Anyway - I like this topic - but - I’m probably going to tailor my existing Raspberry Pi backup scripts to backup my JellyFin server running on Debian 13 (x86_64)… Because I know how they work… They worked on Debian 12 on a Pi4, they currently work on another Pi4, and also worked on an OrangePi 2E+, and also my Pi3 running TVHeadend (Debian 9!).
But when I say “work” I’ve never tested them… it’s more a case of "backing out something that f–ked something else up on Day xx/xx/YYYY
Your backups are only as good as your ability to know how to restore from them… and that’s what DR tests are all about (I hate them!).
I can confirm that . After driving into town and back ( about 30k round trip) , I need an hours sleep before I do anything. I find driving exhausting. Not sure if it is eye fatigue or anxiety or …
i dont have a Harley, but the ag bike or the Landrover or the tractor dont worry me.
Thank you for your well wishes. I think I’m doing OK, at least for now. I’m doing everything my doctor’s have told me to do, and hopefully that will keep things in good order for as long as possible.
As for backups, I agree! That’s why my disaster recovery solution is Rescuezilla. Around the 1st of the month I boot to my Rescuezilla live image and set it up to image my entire Garuda Linux installation, both system and user space. I keep 2 or 3 images handy, deleting the older ones after creating a new one. I’ve never had to recover my system with one of those images and hopefully I never will, but I just think it’s better to have a image backup and never need it than not having it when I do need it.
I also create a file backup of my user’s data with restic. I used the resticprofile app to schedule and set it up the way I want. I think I’ve already posted about it, if you’re interested. I do the snapshot file backup so I can get a copy of any of my files, in the state it was in on any day of the past month, just in case I do something stupid, or inadvertently.
Ernie
Sadly sold my harley a few years back as it was just getting too heavy to control safely and bought a honda monkey instead. Ok cannot take the wife out with me as Its only for one. But I ride with a smile on my face every time, because its a fun bike, no great speed but often get stopped and asked about it.
I am not longer able to give blood, the french will not accept english or better still Yorkshire blood, something to do with mad cows same for organ donations.
I bought my harley when my second wife started divorce procedings, thought if she is going to take everything, I can just ride off into the sun set.
@ernie :
Hi Ernie, ![]()
That is great.
I´ve never dealt with rescuezilla myself, so I cannot say much about it.
But I like the fact that Rescuezilla and Clonezilla are fully interoperable,
meaning that you can use images created by either software with the other.
So backup images made with Clonezilla can be restored using Rescuezilla, and images created with Rescuezilla can be restored using Clonezilla. ![]()
I will definitely look into it.
Seems you can even install rescuezilla as a DEB package on your system:
Well, I think this wouldn´t be the preferred method of backing up your system.
It´s the same with clonezilla. Not a good idea to backup the root partition from a running Linux system.
I guess there are other use cases for installing it.
BTW: I also have installed clonezilla on my Linux Lite system but never used it this way until now. ![]()
My bi-monthly disk image is created from a live USB clonezilla instance, of course.
Many greetings from Rosika ![]()