forum.excito.com

Not wanting to sound too pushy, but is there going to be a fix for the backup to a disk?
As it doesn't work anymore since this update, see my previous replies.

And I don't see anything happening at the moment.

Regards,

Albert

@Albert, we are trying to reproduce it but still it seems as you are the only one experiencing it. We have had one guy on it for two days now but cannot make it fail. That said we can't konw of course, and Carl will contact you to dig deeper into the issue.

albert wrote:Not wanting to sound too pushy, but is there going to be a fix for the backup to a disk?
As it doesn't work anymore since this update, see my previous replies.

And I don't see anything happening at the moment.

Regards,

Albert

Hello Albert,

I've pushed QA to do an extensive retest of the backup system, and they report it works fine after an upgrade. Thus I'm afraid I can't figure out what is the issue with your system.
If possible, could we get access to your B3 to directly see what might be the issue for you?

Hi Carl,

I find it strange that it does not work after the upgrade, before it worked fine. And even newly created backup schemes don't work. What I will do first is a reboot, can't remember I did that after the upgrade, probably not. And see if it works again.

Is there anything I can do to get more info? E.g. do a manual backup on the commandline, and if possible verbose? Shell scripts can be run using a -vx in the first line, don't know if you can do this with perl as well.

Albert

albert wrote:Hi Carl,

I find it strange that it does not work after the upgrade, before it worked fine. And even newly created backup schemes don't work. What I will do first is a reboot, can't remember I did that after the upgrade, probably not. And see if it works again.

Is there anything I can do to get more info? E.g. do a manual backup on the commandline, and if possible verbose? Shell scripts can be run using a -vx in the first line, don't know if you can do this with perl as well.

Albert

Can you confirm you have the latest version of bubba-backend installed. There was a last-minute update to that package, but it happend before we announced this update. But the version of bubba-backend should be 2.5.0.4.

Hi Carl,

I can't check it at the moment (still at work), but will check it tomorrow morning.

I ran an update this morning and nothing came up, so it should be fine.

I may have edited the backup script, not really made changes, but touched it. This may have interferred with the update (file creation date/time is different, maybe even an extra space). Could you do a md5sum on your version of the backup script and put the results here? Then I can check it at home?

Albert

Hi Carl,

Package name Package version
bubba 2.5.0.1
bubba-album 2.5
bubba-backend 2.5.0.4
bubba-frontend 2.5~pre6
bubba3-kernel 1:2.6.39.4-11
filetransferdaemon 0.55
logitechmediaserver 7.7.2.dfsg1-1

root@b3:~# md5sum /usr/lib/web-admin/backup.pl
08b9c0e7ba840966892be50f569b22fb /usr/lib/web-admin/backup.pl

root@b3:~# ls -la /usr/lib/web-admin/backup.pl
-rwxr-xr-x 1 root root 44613 Jun 7 17:03 /usr/lib/web-admin/backup.pl

Hi Carl,

a reboot didn't made any difference.

Albert

Hi Carl,

I mailed this to support as well ([Excito Support #6292]) so you can mail me directly if needed,
but for reference in this thread here are my findings.

I made a copy of the file /usr/lib/web-admin/backup.pl and played around with it.

I have added the following lines to the function find_disk:

foreach my $disk (@$disks) {
foreach my $partition (@{$disk->{"partitions"}}) {
print( "disk_uuid = ", $uuid,"\n" );
print( "uuid = ", $partition->{"uuid"},"\n" );
print( "mountpath = ", $partition->{"mountpath"},"\n" );
print( "dev = ", $partition->{"dev"},"\n" );
print( "\n" );
if($partition->{"uuid"} eq $uuid) {
$retval{"diskfound"} = 1;
$retval{"mountpath"} = $partition->{"mountpath"};
$retval{"dev"} = $partition->{"dev"};
last;
}
}
}
return %retval;


basically the print options to see what is found:

-- output of backup script --

RUN NOW: admin MailWeekly
disk_uuid = 0c09f9ea\-58a8\-497c\-94ce\-d8f1e67ea83b
uuid = 79ea1530-2b1c-4bd9-968b-253c8c9d2c5f
mountpath = /
dev = /dev/sda1

disk_uuid = 0c09f9ea\-58a8\-497c\-94ce\-d8f1e67ea83b
uuid = 39715cbc-c1b9-82a1-0efd-a5235c11aca5
mountpath =
dev = /dev/sda2

disk_uuid = 0c09f9ea\-58a8\-497c\-94ce\-d8f1e67ea83b
uuid =
mountpath =
dev = /dev/sda3

disk_uuid = 0c09f9ea\-58a8\-497c\-94ce\-d8f1e67ea83b
uuid = 0c09f9ea-58a8-497c-94ce-d8f1e67ea83b
mountpath = /media
dev = /dev/sdb1

disk_uuid = 0c09f9ea\-58a8\-497c\-94ce\-d8f1e67ea83b
uuid = 39715cbc-c1b9-82a1-0efd-a5235c11aca5
mountpath =
dev = /dev/sdc1

Backup disk not found, exiting

-- output of backup script --

What I see here, is that the string of disk_uuid contains \ for each -, which compaired to the found uuid doesn't math.
So the error must have been inserted when changing the parsing of the jobdata file.

Kind regards,

Albert

I also have a problem with backup of a special target folder, to a disk via Esata. I can start it, but when I look later, its says not run.

Hi Carl,

I found the problem for the backup issue. When changing the jobdata file to use double quotes arounds strings, and fixing the bug for interpretting this you used the quotemeta function in 2.5.0.1. Unfortunately, quotemeta escapes all non-word characters, including the hyphen. Too bad that the hyphen is used in the disk uuid, so that's where it comes from. So 0c09f9ea-58a8-497c-94ce-d8f1e67ea83b will become 0c09f9ea\-58a8\-497c\-94ce\-d8f1e67ea83b, as you've seen in my previous report in this thread.

Hope you can fix this soon, so I can get the backup running again.

Albert

Apparently NFS v4 server support is still disabled in the B2 kernel. Can anyone from Excito confirm this?

Any news on the backup update?

Apart from that I found another weird error which wasn't there beforer.

If I run update on the commandline (apt-get update; apt-get upgrade) it all works fine, and there are (unfortunately) no updates yet. If I do the same through the webinterface (the update that is), it first loads all the update tables etc, as usual. But at the stage where it should say no updates I get the error "No functional Internet connection was found". This wasn't there before the update 2.5.0.1.

As it works fine on the commandline, I assume something has changed in the webinterface.

Regards,

Albert

The backup fix will release soon.

We cannot reproduce the other error you see, are you sure you had a valid internet connection (to B3) at the time? Please try again.

Hi Johannes,

yes, tried it the past few days. Same error, as said doing it on the commandline works fine.
Any logs I can check specifically (when I get home)?

Albert