Antsle Forum

Welcome to our Antsle community! This forum is to connect all Antsle users to post experiences, make user-generated content available for the entire community and more. 

Please note: This forum is about discussing one specific issue at a time. No generalizations. No judgments. Please check the Forum Rules before posting. If you have specific questions about your Antsle and expect a response from our team directly, please continue to use the appropriate channels (email: [email protected]) so every inquiry is tracked. 

You need to log in to create posts and topics.

Routine Backups for my Antsle

12

Hi all,

I've been trying to come up with a plan to ensure my Antsle VMs (both KVM and LXC) are backed up.   At a rudimentary level, I'm thinking that I need to be able to create an image backup of each virtual disk, and a backup of the config data needed for Antsle to boot and run the VM.

Before I go to work to make something simple and automated to do these backups, I wanted to make sure I wasn't missing anything.  Here's what I'm doing now:

  1. Make a .tgz archive of the entire /etc/libvirt directory
  2. Power off the VMs and take a snapshot giving it a name indicating the date the snapshot was taken... then copy the snapshot something like this:

zfs send antlets/[email protected] | gzip --fast > /hdd/antlets-www-2019-01-12.gz

Then lastly, I ftp these compressed exports and the .tgz of /etc/libvirt off the Antsle to external, removable storage that I take offsite.

So if something awful happened and I had to take a new, unconfigured Antsle and re-install these virtual machines, I know there would be some network configuration to do on the Antsle and some names would need to match.  But assuming I got that top level Antsle configuration right, would the "restore" process consist of: 1) decompressing the snap (pipe the output of gunzip) as input to the ZFS tool to import the exported image as a properly named virtual disk in the zfs /antlets directory, and 2) extracting the contents of  my libvirt tgz into/etc/libvirt?

What am I missing?  Is there a better way to do image backups and restores?

Thanks!

Tim

Greetings Tim,

I use a separate PC (CentOS 7) running the backuppc application.  With the cost of hard drives, you can easily get a couple of terabytes for your backup machine.  It is very possible to use a USB drive on a Raspberry Pi to accomplish the same.

What I would then do is run a second iteration of backuppc offsite to back up the back up system.

Three times in the past ten years I have had to do a complete restore of a system from the backuppc server, and had no trouble doing so.

In using something like this application (it is a wrapper for rsync), you don't need to take down your machines for the backup.

Remember, your mileage may differ.

Jeff

Quote from jeff on January 12, 2019, 10:07 am

Greetings Tim,

I use a separate PC (CentOS 7) running the backuppc application.  With the cost of hard drives, you can easily get a couple of terabytes for your backup machine.  It is very possible to use a USB drive on a Raspberry Pi to accomplish the same.

What I would then do is run a second iteration of backuppc offsite to back up the back up system.

Three times in the past ten years I have had to do a complete restore of a system from the backuppc server, and had no trouble doing so.

In using something like this application (it is a wrapper for rsync), you don't need to take down your machines for the backup.

Remember, your mileage may differ.

Jeff

Thanks, Jeff, and that's a great solution for a file-by-file backup of a VM.  I love rsync (both as a part of the tool you mention and as a great stand-alone method to move lots of data around over the network).  However, I'm more looking for a fast drop-and-go "disaster recovery" solution.  I'm doing "image" backups of the VMs from a snapshot.  I make the snapshot from a powered off state, just to make sure that the VM's file system is properly unmounted and that no fsck or journal replays would be triggered if I restored the image and booted the VM.  The image approach gathers everything - including boot configurations and "partition" layouts - and makes an image at the virtual disk level, as opposed to the VM's internal file system level.  The intent here is to be able to completely restore a VM to a fully configured (and file system error free), boot-and-go state from the image (possibly on a different Antsle)...

The part I'm most in doubt about is... Am I getting everything I need to capture the internal configuration of the Antsle itself?  Is all the VM and Network stuff located in /etc/libvirt?  If I am getting everything, then I can also completely reconstruct my Antsle environment on a new Antsle (in case of a really nasty disaster like a fire) by restoring the configs to the right places (this part IS file-by-file... the .tgz of /etc/libvirt I'm making).

So my question is more "Am I getting everything on the "Antsle config" side that I need to be able to do that kind of "drop and boot" restore to a new, unconfigured Antsle (and/or to my existing one)?"

Thanks for your assistance!

Tim

To make things a little more clear about what I'm doing and what I'm asking...

Here's what I do now...

To back up my Antsle environment for Disaster Recovery purposes using image backup techniques, I do the following:

  1. Power off each Antlet and take a snapshot named as the current date. Then power them back up and proceed to step 2.
  2. Log in to the Antsle with PuTTY and, for each Antlet, as root, run:
    zfs send antlets/[email protected] | gzip --fast > /hdd/AntletName-SnapshotName.gz
    For example: zfs send antlets/[email protected] | gzip --fast > /hdd/antlets-mx-2019-03-10.gz
  3. Create a .tgz of /etc/libvirt ( tar -cvzf libvirt-2019-03-10.tgz /etc/libvirt )
  4. scp (copy) those .gz files and the .tgz of libvirt to a removable drive on my PC (using WinSCP)
  5. Take the removable drive offsite till the next backup cycle.

The question is... would these files allow me to take a fresh out-of-the-box Antsle and fully re-create my running environment by restoring /etc/libvirt and using gunzip and zfs send to restore the Antlets??  What else is needed?  What am I missing?

Thanks!

Tim

Wow - AntMan really needs to program this sort of thing into its user interface.

 

 

*click*   VM backed up into a nice bundle ( to some place from which it can be copied to offline storage !! )

*click*   VM restored from previous backup...

Quote from rockandroller on March 19, 2019, 2:44 pm

Wow - AntMan really needs to program this sort of thing into its user interface.

 

 

*click*   VM backed up into a nice bundle ( to some place from which it can be copied to offline storage !! )

*click*   VM restored from previous backup...

Yes, a 1-click solution in a future version of AntMan would be awesome!   Speaking of which, as a non-business home user, where did the antman upgrade tab go? How do we know about available updates now?  Or are we now no longer entitled to updates!?

 

Tim did you ever get an answer to your question? I need to return my Antsle under warranty due to an addon drive failure. I am looking for an easy way to backup my antlets. I have a 4TB WD Cloud drive in my home office where I would like to store the backup files.

Nope.  They now offer a pay "cloud" service for backing up... but nothing for home users who just want to back up locally... (or to their separately purchased cloud storage).  It's kind of disappointing to see how this product has evolved.  I've never upgraded to edgelinux simply because if I did, and didn't want to purchase a never-ending subscription of some sort, my Antsle would be so crippled up I wouldn't be able to run my home workload any longer... I'd have the capacity, but not the license, to run the VMs I'm running today.  Really, really disappointing.  If it continues, I can always rip out the "Antsle Heart" of my Antsle and install my own open source OS that supports "libvirt" virtual machines and ZFS on the (previously) Antsle hardware.  I still do love the form-factor... no fans, sealed case, no cat hair to vacuum out, low power consumption. 🙂  But support for the non-business customer has pretty much become non-existent.  Really, really too bad.

I also like the hardware (no cat hair FTW!) and I can live with (but not use) a subscription for on-going things (like support or cloud storage, etc), but it is difficult for me to understand subscriptions for "static" features (like Advanced Networking or Clustering).

I'm also trying to figure out what I'm going to do when the hardware does fail.  Even if I have all these backups, when it a drive or memory or something fails, I can't just "open it up" and replace that one part.  I have to send the whole thing back (if it is under warranty).  Do I order 2 and keep the other one on standby?

Even if I were a small business and paying the $29/month for the privilege of using it for a business purpose, what's the process?

Has anyone had a hardware failure on one yet?  A majority of the issues I see in Forum are around O/S or AntMan upgrades.

I did have a drive fail while the unit was under warranty.  At the time (and still perhaps today), had I not been in the unit looking at device status, I'd never have known my mirror set was degraded.  The GUI didn't report any issues.  Again, I never upgraded to EdgeLinux because I didn't want to be limited from using my hardware's full capacity by the crippled up "free" license.  So perhaps hardware failure detection and reporting is better now.  But I digress...

If you were in a business production environment that couldn't tolerate down-time, I'd strongly suggest a spare, an on-site cluster with plenty of spare capacity, or one of their business "fail to cloud" arrangements.  They offer a nice range of options for "high availability"... and that's great for business users... but my particular use case is "cheapskate individual I.T. geek who wants to play at home" and I'm not going to pay a recurring fee for the privilege of using the full capacity of my already-purchased hardware.

With that said, when my Antsle was under warranty, I sent the unit off and had it back in about a week.  The turnaround was fast.  But I was without the Antsle for that time.  After the warranty is up, I don't think there would be any reason not to crack the case and replace any failed components myself.  There is an issue with "putting things back exactly like they were" when you close things up so as to get the thermal coupling you need to dissipate heat using the case as a sink.  But with that caveat, that's the current plan when something in my Antsle breaks, but it has yet to occur.  So far, after the almost "out of box" drive failure, my Antsle has been quite reliable.  The drives and memory are all "commodity" products from what I can tell.  We'll see how things go.

Has anyone here with an out-of-warranty Antsle cracked the case and taken a look inside?  I strongly suspect it's pretty much all "commodity" stuff inside a very unique case.  The motherboard would be the one thing that might be a little unique, but I bet there's a 3rd party manufacturer and part number on it somewhere. 🙂

Good luck!

Tim

12
powered by proof factor - increase conversions with social proof notifications