should the sd card be formatted first?
No, you are dealing with partition images which include the filesystem. Formatting would just be extra writes.
Is that just a given, or is usbimager just magical?
A filesystem is just blocks of data. usbimager writes those blocks of data (sometimes decompressing on the fly).
I validated a full cycle and it worked !
backup
automatically upload to a S3 bucket
download from a S3 bucket
unzip
pishrink
etcher to a sd card
boot up !
Wonder if there is a better process, but AFAIMC it's works and sufficient !
I have been experimenting with the backup plugin for three days now, while following this thread. Here are a few observations:
When you click the "Schedule backup" button in the Backup tab and save your choice (daily, weekly, etc.) you are creating a Schedule Job in the Schedule Jobs tab. If you repeat that process thinking you are changing the job from one schedule to another, you are not. You are creating a separate job. Every time you save a backup schedule in this way you are creating an additional job in the Schedule Jobs tab. If you want to change the backup schedule from daily to weekly, go into Schedule Jobs and delete all but the one you finally choose. The pics below may help for you visual people.
I stumbled across this when I went to burn an image earlier today and found two .ddfull.gz images for June 30. They were created one minute apart, one at 00-00-01, and one at 00-00-02. This may account for the failed/messed up sd cards I burned today.
On two attempts I was not able to get a successful sd card using usbimager on this little Debian box I'm using. The first time when I plugged it into the hc2 and turned it on, the blue light blinked a couple of times and then nothing. Nothing via command line, and nothing via the browser. The second attempt using the other .gz image with usbimager I was able to boot the hc2 but the login screen failed to allow a login. I was able to login to the command line. Next I burned a sd card using Etcher on my Mac and was successful. Sort of. I was able to log in. I clicked around the tabs and everything looked fine until I looked at the Shared Folders tab and SMB tab. My Media folder was missing. I had just created it a few days ago in preparation for testing Jellyfin. The Media folder was absent from the shares on both my Debian machine and my Mac. But Jellyfin still worked when I went to the web page. Files in the Media folder showed up and played fine.
Now that I have all jobs deleted except one daily job in the Schedule Jobs tab, I will try to create a successful sd card with tonight's backup.
I will report in with an update tomorrow.
Do you understand the purpose of other files after backup?
No.
I was going to test last night’s backup today and a thunderstorm came through and knocked our power out. At least I got a chance to test my UPS plugin:
No.
I was going to test last night’s backup today and a thunderstorm came through and knocked our power out. At least I got a chance to test my UPS plugin:
When I installed my new UPS and the UPS plugin, I tested it by pulling the power from the UPS mains eheh 🤣
Since we're talking about backup plugin, is there a way to save the backup on a shared folder as I do already (it's another SSD) and also save it to a cloud like Google Drive? I know I can use rclone and save the backup directly on the mounted cloud drive, but I also would like to retain a local copy in case internet goes down.
Do you understand the purpose of other files after backup?
I have explained this many times and the extension should give a clue what it is but why not one more time...
*.fdisk is the output of fdisk -l. Informative if you wanted to see the disk layout at the time of backup
*.blkid is the output of blkid. Informative if you wanted to see the filesystem list and uuids at the time of backup
*.grub is the first 446 bytes of the disk where the grub setup is
*.grubparts is the first 512 bytes of the disk where the grub setup and partition table are
*.packages is the output of dpkg -l to show all the packages installed on the system at the time of backup
*.ddfull is the backup
Since we're talking about backup plugin, is there a way to save the backup on a shared folder as I do already (it's another SSD) and also save it to a cloud like Google Drive? I know I can use rclone and save the backup directly on the mounted cloud drive, but I also would like to retain a local copy in case internet goes down.
Simplest solution would be to run a scheduled job to synch the local shared folder with the cloud folder.
Simplest solution would be to run a scheduled job to synch the local shared folder with the cloud folder.
It's so simple I haven't though about it. Thanks, will work on it!
Simplest solution would be to run a scheduled job to synch the local shared folder with the cloud folder.
That is basically what I did:
As the plugin removes x days old files, the subsequent Rsync job does the same. I can then download the gz file to desktop and burn the image using usbimager. It works great/fast on my Debian machine. Can’t seem to get the permissions right to get it to burn on my Mac. Trying to move to Linux anyway. 
Alles anzeigen*.fdisk is the output of fdisk -l. Informative if you wanted to see the disk layout at the time of backup
*.blkid is the output of blkid. Informative if you wanted to see the filesystem list and uuids at the time of backup
*.grub is the first 446 bytes of the disk where the grub setup is
*.grubparts is the first 512 bytes of the disk where the grub setup and partition table are
*.packages is the output of dpkg -l to show all the packages installed on the system at the time of backup
*.ddfull is the backup
I’m sorry. I’m slow/I missed that. I have pasted that into my little how-to notes. Thanks so much. Apologies also for pulling a bit off topic above, but auanasgheps i was going to do what you did (with UPS) but got busy and never did. I just crossed my fingers.
Alles anzeigenI don't think any OMV backup plugin does what I need.
Here is the script. I run it daily on cron @ 3:00am. It takes under eight minutes to backup the OMV 16GB SSD.
You will have to change it to identify your OMV system disk, identify your OMV major version number, and identify the drive and directory where to store the image file.
BashAlles anzeigen#!/bin/bash # Script to make dd image of system drive now=$(date +"%Y.%m.%d.%H.%M.%S") file="omv-5-$now.img" cd /srv/dev-disk-by-label-d1/omv-backup-imgs dd if=/dev/disk/by-id/usb-SAMSUNG_SSD_UM410_Series_100000000000-0:0 of=$file bs=1M
Sorry to get in the middle of this thread i tried your script it is working perfectly from CLI, I can boot from the created image with etcher, but I cannot make it work in cron from the gui by imputing it in the command field.
I saw this
RE: Error if I run my script from "scheduled jobs"
but it did not work for me.
What am I missing?
Below are my values for this code
#!/bin/bash
# Script to make dd image of system drive
now=$(date +"%Y.%m.%d.%H.%M.%S")
file="omv-5.5.3-1-$now.img"
cd /srv/dev-disk-by-label-ssd-nvme/omv-backup-imgs
dd if=/dev/disk/by-id/usb-SanDisk_Cruzer_Glide_3.0_4C530001171010104183-0:0 of=$file bs=1MEdit
Also I had no problem to find dev/disk/by-id/ for the USB stick from my intel server but how to find it from a SD card from my rockpro?
Not sure what to tell you about not being able to run it in the scheduler. Make sure you have the full path name to the script file in there and that the script is chmod to be executable.
Use this command to fully identify all disks on a system:
tree /dev/disk -a
I tried this for my intel
ls -l /dev/disk/by-id/usb*
or
ls /dev/disk/by-id/
but
root@openmediavault:~# tree /dev/disk -a
-bash: tree: command not found
At least I found it it was /dev/disk/by-id/mmc-SN64G_0xd0e40f12 now just need to figure out how run from the scheduler.
root@openmediavault:~# tree /dev/disk -a
-bash: tree: command not found
sudo apt-get install tree
root@rockpro64:~# tree /dev/disk -a
/dev/disk
|-- by-id
|-- mmc-SN64G_0xd0e40f12 -> ../../mmcblk0
Thanks
Got it working in cron job, of course it was some errors from me.
Thanks again
Now what could be nice maybe in the same or in a separate rsync script from dev-disk-by-label-d1 for the file omv-5.img to a shared folder on the windows PC?
You can try sometime I got errors in etcher after verification and the card is still booting, but a function in OMV may not work as expected.
Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!
