https://www.danube.cloud | https://github.com/erigones/esdc-ce/wiki/Release-Notes
@toybomb:matrix.org i had the same question a while ago. yanchii https://gist.github.com/cnrd/9b8d50a17a3f7860298337bbbb084d12
@toybomb:matrix.org i had the same question a while ago. yanchii mentioned that most scripts actually require the other side to have zfs. this rclone solution doesn't but isn't very well structured https://gist.github.com/cnrd/9b8d50a17a3f7860298337bbbb084d12
- one is that when you build your own cloud, it makes sense to have backups managed inside it... even offsite backups can be managed by DC. Example: install a remote DC node somewhere elsewhere and push backups there.
- second is that you want to have backups outside the current orchestration system. The easiest way is to utilize
filebackups (instead of zfs send/receive) using DC and then have these files uploaded by some cron script anywhere you like.
esbackup tool https://github.com/erigones/esdc-ce/blob/master/bin/esbackup
esbackup with required parameters during VM backups
And the High availabilty of management VMs?
That's a different story. But now I don't know what exactly do you mean in connection to HN/CN. Same migration rules apply.
you have several options before choosing mgmt HA because it is the most complicated option and it's not maintentance-free.
setup periodic backup of admin VMs and if your HN breaks inrecoverably , follow the DC restore procedure: https://docs.danube.cloud/user-guide/howto/disaster-recovery.html#recovery-without-danube-cloud-api-or-gui
setup VM replication for
mgmt01and if the HN disappears, just start the destinationmgmt01VM on the destination nodesetup HA. There are ansible, python and shell scripts to do that. But you will have corosync/pacemaker postgresql cluster and you should know how to recover it in case something goes wrong... the same as with any other database cluster.
62 replies
Hi Everybody
Today I tried to install the new LX image ubuntu 20.04 from https://github.com/omniosorg/lx-images
And it worked without any modification with DC 4.3.
These are the steps I used:
I have created a new LX server with DC using the latest ubuntu LX available 16.04
I copied the <UUID> of the created zone
I stopped the zone with DC
I went to the CN node via terminal
I created a working folder under /zones
mkdir /zones/WORK
cd /zones/WORK
I Moved the existing /zones/<UUID>/root folder in /zones/WORK/rootOLD
mvdir /zones/<UUID>/root /zones/WORK/rootOLD
I downloaded the omnios LX ubuntu 20.04 available in the github site:
wget --no-check-certificate https://github.com/omniosorg/lx-images/releases/download/2020-10-05_16-35-52/lx-ubuntu-20.04-2020-10-05_16-35-52.tar.xz
mkdir /zones/WORK/root
xz -d lx-ubuntu-20.04-2020-10-05_16-35-52.tar.xz
tar xvf lx-ubuntu-20.04-2020-10-05_16-35-52.tar -C root
mvdir root /zones/<UUID>/root
At this point I restarted the zone from DC and it booted correctly:
[root@node (DC) ~]# zlogin 16bd71ff-8dd4-43bf-b556-58723a289b90
[Connected to zone '16bd71ff-8dd4-43bf-b556-58723a289b90' pts/2]
Welcome to Ubuntu 20.04.1 LTS (GNU/Linux 4.4 x86_64)
Documentation: https://help.ubuntu.com
Management: https://landscape.canonical.com
Support: https://ubuntu.com/advantage
Last login: Thu Oct 22 07:13:57 UTC 2020 from zone:global on pts/4
root@16bd71ff-8dd4-43bf-b556-58723a289b90:~# ls
root@16bd71ff-8dd4-43bf-b556-58723a289b90:~# uname -a
Linux 16bd71ff-8dd4-43bf-b556-58723a289b90 4.4 BrandZ virtual linux x86_64 x86_64 x86_64 GNU/Linux
otis * void linux seems to have some "good" concepts
toybomb But void may actually allow me to run mailinabox by manually setting package versions
toybomb on the ubuntu 18.04 and 20.x images
toybomb * on the ubuntu 18.04 and 20.x lx zones
toybomb i spent a few days trying to get some services running last month when tried those images
toybomb but devuan might really be a good idea