Just wanted to feedback a small issue while restoring in our migration:
Starting restore of 1 domains from local file /root/backup-stat-final-2012-12-11/stat.example.com.tar.gz .. Extracting backup archive file .. .. done Re-creating virtual server stat.joomlapolis.com .. • Creating administration group stat .. • .. done Creating administration user stat .. .. done Creating aliases for administration user .. .. done Adding administration user to groups .. .. done Creating home directory .. .. done Creating mailbox for administration user .. .. done Adding to email domains list .. .. done Adding default mail aliases .. .. done Adding new virtual website .. .. done Adding webserver user www-data to server's group .. .. done Performing other Apache configuration ..
root@mu32:~# ps aux | grep /usr/share/webm
root 4021 0.0 0.0 78880 540 ? Ss 00:19 0:00 /usr/share/webmin/virtual-server/lookup-domain-daemon.pl
root 4055 0.0 0.0 123668 1692 ? Ss 00:20 0:00 /usr/bin/perl /usr/share/webmin/miniserv.pl /etc/webmin/miniserv.conf
root 5459 54.1 97.5 21501920 14962736 ? R 00:49 1:20 /usr/share/webmin/virtual-server/restore.cgi
Redoing operation once the process died (without output in web window more than above), it worked.
So nothing major, but it can eat memory as much as available (RAM+swap...).
Comments
Submitted by JamieCameron on Thu, 12/13/2012 - 11:31 Comment #1
So it hung at the "Performing other Apache configuration" step?
Submitted by beat on Thu, 12/13/2012 - 12:18 Comment #2
Yes. (sorry for the bad formatting above, copy-pasted it from a textpad).
Btw: with very latest Ubuntu 12.04LTS up to date, and very latest Virtualmin Pro releases (same versions, just different servers, for the other bug repots, 1 of today, and 1 of a few days back).
Might have some more bug reports (at least one with RNDC) coming, but need to sort out the reasons first, as we are migrating all our servers these days.
Submitted by JamieCameron on Thu, 12/13/2012 - 12:23 Comment #3
If you create a new domain on the same system (not from a restore), does it hang in the same way?
Submitted by beat on Thu, 12/13/2012 - 14:03 Comment #4
No, 2nd domain restore (and also 2nd trial of first one) went well...
So it's not critical here, just wanted to give a heads up in case you hear something similar elsewhere it might help.
Submitted by JamieCameron on Thu, 12/13/2012 - 15:28 Comment #5
Ok, thanks ..
I think to debug this further I'd need to see the backup file that the failed restore was from.
Submitted by beat on Fri, 12/14/2012 - 05:14 Comment #6
Hi Jamie, It's a huge database (5 Gigs) so guess you don't want that in your email box! LOL
But I had to move the site, and have quite some more on this:
1st bug:
I had to move that virtual server to another Virtualmin Pro installation, so backed up all features (except virtualmin general settings) for that site and another one, using the new separate files in folder.
That site had: - Private IP address - valid SSL certificate
When I restored to another server with Virtualmin Pro containing already one virtual server with another private address and ssl certificate, I had exactly same problem. I selected: Keep original IP address. This happened Repetitively (as I had to kill that 12 Gigs process, i had to remove user, group and non-virtual server in Webmin / Servers / Apache: (BTW WOW!!!: checkmark a virtual server, click delete button: GONE, no warning "are you sure?" question asked!!!)
(calling that non-confirmation bug 2)
Interestingly, the SSL server wasn't created. So I guess that bug is at creating The SSL port 443 Apache server with private IP address that didn't exist on the system. I looked in networking, and the private address eth0:x interface wasn't yet created.
Maybe bug is simply that it tries creating Apache non-SSL 80 port and SSL 443 port and while it strangely succeeds on the 80 port it is looking with a memory-leak to the IP address interface for the SSL port ?
To make it work (i recalled that now for the first server bug too): I had to first create the Virtual Server with the features wanted (correct private IP, HTTPS), then restore the server. Here the "features that will be restored" from that backup:
Virtual server password, description and other detailsAdministration user's password, quota and Cron jobs
Mail/FTP users and mail aliases
Apache webserver configuration
Apache SSL webserver configuration and certificate
Logrotate configuration for log file
Contents of server's MySQL databases
Webmin ACL files
Server's home directory and web pages
Changung title of the issue from: "First virtual site restored from a Ubuntu 8.04LTS backup on a new server Ubuntu 12.04LTS with Piwik script ate up 21 GB of RAM" to: "Site restore from backup with "Original Private IP address" (not existant) + SSL vert fails and eats up 21 GB of RAM"
3rd bug:
Also: you probably noticed that the backup options: Last field of "Backup Virtual Servers" is garbage in latest Virtualmin pro with 3 (radio) with only last one with text (new blank Vmin pro install too):
Backup level (radio) (radio) (radio) Neither (all files, and don't update incremental state)Submitted by JamieCameron on Fri, 12/14/2012 - 11:48 Comment #7
Ok, I see the bug now .. this can be triggered only when restoring a domain with SSL enabled.
We are going to release a minor update tomorrow that will fix this, and the backup level bug you mentioned.
Submitted by beat on Fri, 12/14/2012 - 12:54 Comment #8
Great! Saturday release is Perfect timing ! Having quite some sites to move to a new host by backup and restore on Sunday :-) Thank you Jamie!
Submitted by Issues on Fri, 12/28/2012 - 12:58 Comment #9
Automatically closed -- issue fixed for 2 weeks with no activity.