Disk usage for failed jobs

[Christer Fletcher]

Hello all

I wanted to ask what happens when a job fails. I'm doing backup of my wife's laptop which she turns on and off as she uses it. It's no problem doing Incr of Diff jobs but the Full job often does not complete during the window that her laptop is on.

When a full job fails, will it purge the data it wrote to the SD, will that data be reused for the next job or is it just dead bytes wasting space?

I also wanted to ask if this following is a sane configuration to handle computers that comes on/off randomly. I start a new job every night.

Job {
...
Rerun Failed Levels = yes
Reschedule On Error = yes
Reschedule Times = 9999
Allow Duplicate Jobs = no
Cancel Lower Level Duplicates = yes
}

Best Regards
Christer Fletcher

[Philipp Storz]

Hello Christer,


Am 03.08.2013 22:11, schrieb Christer Fletcher:> Hello all

>
> I wanted to ask what happens when a job fails. I'm doing backup of my wife's laptop which she
turns on and off as she uses it.

It's no problem doing Incr of Diff jobs but the Full job often does not complete during the window
that her laptop is on.

If you installed the traymonitor, she can easily see that the backup is working so should not switch
off the computer.

>
> When a full job fails, will it purge the data it wrote to the SD, will that data be reused for the
next job or is it

just dead bytes wasting space?

The data will not be purged and not be reused by a next job. In the end its a waste of space.
If really needed, the data wrote until stopped can be recovered by doing some tricks.

>
> I also wanted to ask if this following is a sane configuration to handle computers that comes
on/off randomly.

I start a new job every night.
>
> Job {
> ...
> Rerun Failed Levels = yes
> Reschedule On Error = yes
> Reschedule Times = 9999
> Allow Duplicate Jobs = no
> Cancel Lower Level Duplicates = yes
> }
>
> Best Regards
> Christer Fletcher

Hm, "Reschedule Times = 9999"
seems a bit too big.

What is your reschedule interval?
If jobs in your queue are piling up that is not really a good idea.

I think the best idea in your case is that you show your wife that if the traymon is blinking, she
should not shutdown the computer.

What should be technically possible is to implement a "resume" command that can resume a broken
backup and only backup what was not backed up util now.

Best regards,

Philipp





--
Mit freundlichen Grüßen

Philipp Storz philip...@bareos.com
Bareos GmbH & Co. KG Phone: Phone: +49221630693-92
http://www.bareos.com Fax: +49221630693-10

Sitz der Gesellschaft: Köln | Amtsgericht Köln: HRA 29646
Geschäftsführer: Stephan Dühr, M. Außendorf,
J. Steffens, P. Storz, M. v. Wieringen

[Ian Rawlings]

On Saturday, 3 August 2013 21:11:00 UTC+1, Christer Fletcher wrote:
> Hello all
>
> I wanted to ask what happens when a job fails. I'm doing backup of my wife's laptop which she turns on and off as she uses it. It's no problem doing Incr of Diff jobs but the Full job often does not complete during the window that her laptop is on.

Laptops are a problem and it's not always ideal to have to leave it on particularly if the user doesn't feel like it's their problem - no idea if your wife is one of those but it's a problem with kids, wives and management staff (hehe).

So what I do is to use rsync to back up to a server disc, rsync can tolerate going offline and will catch up on the next job. I then do a backup of the laptop directory. For a home environment this is easy, not so easy in a business. I don't know if you've already considered this, if not then it's worth thinking about.

Thanks,

Ian Rawlings

[Christer Fletcher]

My wife has the tray monitor installed but I'm with Ian on this that it should not be her problem, it should just work. And it's indeed not always beneficial to leave the laptop running for 2 more hours if it's not used otherwise, neither for power bill or environment.

About the job queue it's no real issue as I have the following config parameters so the previous or new job is always canceled each time a new job starts so there's always just one job for each client in the queue. It does generate a fair amount of canceled jobs though that clutters the message list a bit.

About rsync I might give it a try, it would use quite a lot of extra disk space on the server though so I'm not a fan of it. Also, I earlier used backuppc with rsyncd transer and I never achieved very good speeds with it, bareos gives me much better speeds.

I'm ordered a new router so I'll see if it will finish the full jobs normally later as I should get better speeds with new router.

It would be awesome if bareos could re-use partial jobs, and if not, at least purge the data from storage if it's not being used either way.

Thanks for your support
Regars
Christer Fletcher

[Ian Rawlings]

On Monday, 5 August 2013 22:43:53 UTC+1, Christer Fletcher wrote:
> About rsync I might give it a try, it would use quite a lot of extra disk space on the server though so I'm not a fan of it. Also, I earlier used backuppc with rsyncd transer and I never achieved very good speeds with it, bareos gives me much better speeds.

I use an rsync daemon on a file server, no encryption but it is password protected, that helps speed things up a bit. Using the --partial flag on the client means it can even pick things up half-way through a file after an interruption. I've generally found speed to be good, even backing up my laptop with large virtual machine disc files doesn't take long. I am however connecting via wired ethernet, not wireless, which helps.

Thanks,

Ian Rawlings

[Christer Fletcher]

If I add spooling (let's stay 20GB of spooling space) and my dear wife turns of the laptop after it's done backup of lets say 15GB, will thouse 15GB be despooled or discarded as the backup job exits with error?

Regards
Fletcher

[Marco van Wieringen]

If a Job fails it will discard the spool data so if you set the spool
size bigger then the actual data it will discard the data and only your
catalog will have a failed job entry in the backup catalog. When your data
is bigger then the set spool size it will despool the data during the
backup and you still end up with x Gb of data of the failed job.

--
Marco van Wieringen marco.van...@bareos.com
Bareos GmbH & Co. KG Phone: +49-221-63069389
http://www.bareos.com

Sitz der Gesellschaft: Köln | Amtsgericht Köln: HRA 29646

Komplementär: Bareos Verwaltungs-GmbH