Think this may be a bug as I have carried out a clean install of the OS (CentOS 7.5) and Virtualmin (with no errors reported) and unfortunately I am still experiencing the same issue. Both servers are pretty much identical, so I am unsure what is going on.
I've been unable to restore either via the virtualmin login webpage, or via command line with the following command
virtualmin restore-domain --source /home/backups-for-restore/mysite.com.tar.gz --all-features (which reports:
No domains to restore specified)
Looking on GH it seems the command line tool throws that error when there is no
--source parameter (https://github.com/virtualmin/virtualmin-gpl/blob/master/restore-domain....) but it's there.
Is this a bug perhaps? Or am I doing something wrong?
Submitted by Brook on Sat, 08/11/2018 - 22:20 Comment #1
Looks like I was missing this
--domain mydomain.com. The web interface is still not working though so I'll leave this open.
Submitted by JamieCameron on Sat, 08/11/2018 - 22:53 Comment #2
What goes wrong with the web interface exactly?
Submitted by Brook on Sat, 08/11/2018 - 22:58 Comment #3
Sorry Jamie I forgot to link to the forum thread: https://www.virtualmin.com/node/58406
Basically it hangs on
Show what will be restored.
Submitted by JamieCameron on Sun, 08/12/2018 - 00:41 Comment #4
Did you get any error messages in
/var/webmin/miniserv.errorwhen you tried to restore?
Submitted by andreychek on Sun, 08/12/2018 - 08:44 Comment #5
How large is the backup file being restored out of curiosity?
Submitted by Brook on Sun, 08/12/2018 - 12:24 Comment #6
JamieCameron - the only thing of note was a timeout message:
There are however lots of these:
Andreychek - it was 43mb, another I tried was 35mb - both had failed, but both worked successfully via the command line.
I have used the webUI for restores on older servers and they have always worked well (and fast).. so not sure what is causing this to hang tbh.
Submitted by JamieCameron on Sun, 08/12/2018 - 16:28 Comment #7
Those "Timeout : Waited more than 60 seconds for request data" errors look suspicious. Are you accessing your virtualmin server directly, or via some kind of proxy ?
Submitted by Brook on Sun, 08/12/2018 - 16:34 Comment #8
Directly - the only thing I've changed is the port.
(Things like running a cron via the web UI seem fine too.)
Submitted by JamieCameron on Sun, 08/12/2018 - 17:05 Comment #9
Ok .. and which browser are you using?
Submitted by Brook on Sun, 08/12/2018 - 17:14 Comment #10
Safari Version 11.1.1 (13605.2.8) on MacOS.
Btw, I'm not sure if it's related, but I have encountered one failure on restoring via command line - I'm not too bothered by this one as I don't mind decommissioning this particular site, but here's the results if it's any interest to you:
The reason the sub domain is called 'directory' is because the site is a Directory (that lists things). Just in case you were wondering, haha!
(PS, I'm not sure if Markdown code blocks are working properly here, I usually do ```text for text blocks and they usually display correctly on other forums which use Markdown)
Submitted by Brook on Sun, 08/12/2018 - 17:24 Comment #11
Oh dear, the same error has happened on another restore - unfortunately this time I cannot discard this site.
Any ideas what might be causing this?
Here are the last few lines from the logs (doesn't seem relevant, but just in case....)
Submitted by Brook on Sun, 08/12/2018 - 17:49 Comment #12
Just a quick note to add that both of these sites are Ruby on Rails sites - and from what I remember, the instructions to install
mod_ruby(i.e. Passenger Fusion) has changed to the following:
Submitted by Brook on Sun, 08/12/2018 - 18:34 Comment #13
Could this be triggering those errors...
In my Ruby apps my directory structures are like this:
name-of-appis repeated - could that be why the error is saying:
Submitted by Brook on Sun, 08/12/2018 - 18:46 Comment #14
Yep that was it!!!!
I changed the directory name, so the path is now:
And the restore worked without errors :D
(Just as a note though, this wasn't an issue last time I carried out a restore - probably over a year ago now tho).