How long should a restore from backup take?

Use Samba, copy /config including all hidden files and subdirectories to a pc. Restore by copying them back. Takes less then a minute. (OK, only HA, not the addons)

unfortunatley I needed to replace my SSD, so I thought just restore a backup, but several attempts and still nothing
I installed Ssh and checked the ā€œtopā€ command,
I noticed the nic usages is really high, even before starting the restore.

Mem: 1207612K used, 2721204K free, 1224K shrd, 68084K buff, 555552K cached
CPU:   0% usr   0% sys   0% nic  98% idle   0% io   0% irq   0% sirq
Load average: 0.01 0.16 0.12 2/398 472
  PID  PPID USER     STAT   VSZ %VSZ CPU %CPU COMMAND
  423   357 root     R     1708   0%   0   0% top
  246    25 root     S     7000   0%   2   0% ttyd -d1 -i hassio -p 64296 tmux -u new -A -s homeassistant zsh -l
  356   354 hassio   S     6496   0%   1   0% sshd: hassio@pts/0
  426   424 hassio   S     6496   0%   1   0% sshd: hassio@pts/1
  424   325 root     S     6252   0%   0   0% sshd: hassio [priv]
  354   325 root     S     6252   0%   1   0% sshd: hassio [priv]
  325    27 root     S     6244   0%   3   0% sshd: /usr/sbin/sshd -D -e [listener] 0 of 10-100 startups
  357   356 root     S     5524   0%   3   0% -zsh
  427   426 root     S     5404   0%   3   0% -zsh
   27     1 root     S      204   0%   3   0% s6-supervise sshd
   26     1 root     S      204   0%   0   0% s6-supervise s6rc-oneshot-runner
   24     1 root     S      204   0%   2   0% s6-supervise s6rc-fdholder
   15     1 root     S      204   0%   0   0% s6-supervise s6-linux-init-shutdownd
   25     1 root     S      204   0%   1   0% s6-supervise ttyd
    1     0 root     S      200   0%   3   0% /package/admin/s6/command/s6-svscan -d4 -- /run/service
   17    15 root     S      192   0%   1   0% /package/admin/s6-linux-init/command/s6-linux-init-shutdownd -c /run/s6/basedir -g 3000 -C -B
   35    26 root     S      180   0%   0   0% /package/admin/s6/command/s6-ipcserverd -1 -- /package/admin/s6/command/s6-ipcserver-access -v0 -E -l0 -i data/rules

then after I click restore,
the interface looses connection, but ssh stays alive so HAOS doesnā€™t restart.

But Load increases

Mem: 2312596K used, 1616220K free, 1132K shrd, 68372K buff, 1821188K cached
CPU:   6% usr  19% sys   0% nic  74% idle   0% io   0% irq   0% sirq
Load average: 0.78 0.32 0.17 4/363 524
  PID  PPID USER     STAT   VSZ %VSZ CPU %CPU COMMAND
  246    25 root     S     7000   0%   2   0% ttyd -d1 -i hassio -p 64296 tmux -u new -A -s homeassistant zsh -l
  356   354 hassio   S     6532   0%   3   0% sshd: hassio@pts/0
  426   424 hassio   S     6496   0%   1   0% sshd: hassio@pts/1
  424   325 root     S     6252   0%   0   0% sshd: hassio [priv]
  354   325 root     S     6252   0%   1   0% sshd: hassio [priv]
  325    27 root     S     6244   0%   3   0% sshd: /usr/sbin/sshd -D -e [listener] 0 of 10-100 startups
  357   356 root     S     5532   0%   3   0% -zsh
  427   426 root     S     5404   0%   3   0% -zsh
  524   357 root     R     1708   0%   1   0% top
   27     1 root     S      204   0%   3   0% s6-supervise sshd
   26     1 root     S      204   0%   0   0% s6-supervise s6rc-oneshot-runner
   24     1 root     S      204   0%   2   0% s6-supervise s6rc-fdholder
   15     1 root     S      204   0%   0   0% s6-supervise s6-linux-init-shutdownd
   25     1 root     S      204   0%   1   0% s6-supervise ttyd
    1     0 root     S      200   0%   3   0% /package/admin/s6/command/s6-svscan -d4 -- /run/service
   17    15 root     S      192   0%   1   0% /package/admin/s6-linux-init/command/s6-linux-init-shutdownd -c /run/s6/basedir -g 3000 -C -B
   35    26 root     S      180   0%   0   0% /package/admin/s6/command/s6-ipcserverd -1 -- /package/admin/s6/command/s6-ipcserver-access -v0 -E -l0 -i data/rules

So itā€™s doing something,
only no clue what it does,

Nic usage goes down a bit,

I think ā€œnicā€ stands for ā€œniceā€ and I donā€™t know what it means, I first thought ā€œnetwork interface cardā€, but now I think itā€™s just nice
.

I definitely think the restore backup process could use a little love from the developers, But I also know that there is plenty of work which can use some.

Agreed. On my NUC with i7 restore takes as long as 3-4 hours, while CPU doing basically nothing (load is a few percent).

I recently decided to switch from a cheap 128GB SSD back to a more expensive 256GB SD card and I gave up using the restore function. I just canā€™t wait that long.

I simply use balenaEtcher to copy the whole disk. It was quicker and always work.

1 Like

Yeah, well, thatā€™s possible in your case, but not if something goes wrong. In this case only restore helps, or even worse: reformat SSD, reinstall HA and only then do a restore.
I really wonder what on earth takes sooooo loooongā€¦ ??

I just spent the last couple hours getting to the point of a restore onto a newly upgraded Samsung EVO SSD. I then hit this ā€œI wonder how long this will takeā€ point after seeing the restore without a progress indicator. Googling/searching, I found the comments above.

I abandoned the restore and took the advice of @starryalley above: took everything back a few steps to a basic balenaEtcher to clone and Iā€™m now up and running 30 mins later. Everything looks okay and I was able to test a kitchen switch, but will spend a little more time validating the restore. So far it appears okay and Iā€™m now seeing the bigger available disk space. I thought I would have to go back and do some manual additional steps to expand a drive or something, but it appears to be okay.

Talk about easy! Great tip. Skip the backup and restore from backup!

Restoring from an NUC to an Unraid VM - Backup is about ~750mb

Running on a N6005 Unraid VM since ~60 minutes. Fingers crossed that this will work. Seems like itā€™s only utilising 1 core out of 4 from the VM while restoring.

Please upvote this feature request:

7 Likes

Voted for that feature request. My HA installation went somehow broken after trying to update to 2023.9.3, HA never woke up after update. Then the basic turn power off - turn power on didnā€™t help either, said my HA IP address was unreachable. So, since nothing else to do, I did that classic balenaEtcher to my SSD (I have RP4 with SSD), and then at least I got life to my HA IP address again with that omboarding process. Uploaded my full backup from last night (1,9 GB) and itā€™s still restoring after one hour. Seems like I need to wait at least 2 hours moreā€¦ But giving us some info, if not progress bar, at least something else than forever running circleā€¦

OS 11 broke one of my HACS integrations so trying to roll back with the latest backup. 6.5 gig. Now 4.5 hours and counting!! Fingers crossedā€¦

Edit:
Took 8 hours and then realizing that the backup of course doesnā€™t include the os :sweat_smile::grinning:

So after 8 hours spent for nothing and 5 mins to downgrade the os (from cli, ha os update --version 10.5) everything is fine!

1 Like

Which integration was broken?

Hassio plejd.

1 Like

After 4 hours I gave up restoring a complete 750MB backup to a sd-card on an Odroid N2+.
A bit more information about the restoring process would be greatly appreciated (already voted for the progress indicator feature request).

1 Like

I think you gave up just before endā€¦ :sleepy:
Although i support progress indicator iā€™d rather vote for solving this slow process and have faster restoreā€¦ I can confirm that itā€™s definitely not machine problem, since restore took a good 4 hours on i7 6x series NUC with 32GB of ram, running HAOS.

It is a process problem. It takes far less time to restore manually the core component.

If you search Github for one of the issues which was created regarding the slow restore, then you will find a description how to manually restore core in 5 minutes to a working order.

ahaā€¦ sound great, butā€¦ thereā€™s over 2000 ā€œcasesā€ for core aloneā€¦ do you have some links by any chance?

If you think using the search for me is easier than for you, then you are absolutely wrong! And you should do a favour!

1 Like

Well, that depends on your and my knowledge of english, resulting in finding correct words for searchingā€¦

Thanks for the link, although itā€™s more for cases when HA doesnā€™t work anymore, not for cases when it works, but a person wants to go back for various reasons (like one integration not working after update). And itā€™s hardly 5 minutes of workā€¦ just reinstall HA takes more than that. But, itā€™s less than 4 hours, that part is true.

You must be a mathematician!

You can restore core by reverting the image to an older one (which you wanted to restore) and restoring the files through the ssh addon from the backup. You donā€™t have to reinstall HA.

But all it is an overcomplicated manual process which I do not advise, unless you really know what you are doing.

Until the issue is fixed. You should stay with restoring backups the official way!

1 Like

Yeah, luckily it doesnā€™t happen often. Although i never reverted HA manually i did ā€œgo backā€ once or twice ā€œofficialā€ way, i just ran restore before going to bed and in the morning all was ok. I do some basic linux stuff every now and then, but thatā€™s all. I agree, itā€™s complicated and easy to do even more damage.
(and, sorry, not mathematician, true electronicystā€¦)