Config hint for backup / bareos rookie

[Mi Zi]

hello all

i am a "bareos" beginner and have to get the following request up and running under time pressure:

Requirement:

Full backup from 3 Windows Server to external LTO-8 Tapedrive (total of approx. 5 TB of data)

the Full backup should run weekly for all Servers stored into one Tape

No Differential or Incremential only Fullbackup every week.

and at present time 6 Tapes are available

currently i have a "clean" bareos installation up and running (20.0.3)

the external Tape drive is connected and i did some tests and it seems working

what should the pool / volume config look like ... as well as the storage device config!

thx in advance

[Brock Palen]

I only use clients on Windows so have never ran the director or storage on Windows.

By tape drive working I’m assuming you mean you used btape and ran the tests included with it? If so you should have the approrate ’storage’ config and the storage director working. If not here is my config for my stand alone tape drive, highly recomend use spool if your spool location in your server is fast enough to avoid drive start / stops, for LTO8 that probably means something that can sustain 200MB/s+ preferablly closer to 500MB/s but all LTO8 drive have speed matching. Spooling will avoid a lot of start/stops.

# goes in bareos-sd.conf
Device {
Name = T-LTO4
Autochanger = no
Drive Index = 0
Media Type = LTO4
Archive Device = /dev/nst0
Device Type = Tape
Maximum File Size = 20000000000 # 20 GB
Spool Directory = /mnt/spool/Q-LTO4
Maximum Job Spool Size = 80000000000
Maximum Spool Size = 160000000000
Drive Crypto Enabled = Yes
Query Crypto Status = yes
Maximum Concurrent Jobs = 1
}


# storage config goes in the director config and says how to talk to the devices on the storage deamon
Storage {
Name = T-LTO4
Address = myth
Password = “<snip>"
Device = T-LTO4
Media Type = LTO4
Maximum Concurrent Jobs = 1
Auto Changer = no
}

# pool config goes in director config and says how to treat volumes. You don’t need “next pool” that’s used for migraiton /copy jobs. For your use case you probably want to adjust Volume Use Duraction so the next week the system asks for a new volume (tape) so your not stacking jobs across multiple weeks on one tape. This is all your volume config also. Notice how it attached the volumes in this “pool” to the storage deff above which tells how to connect to teh stroage server etc.
Pool {
Name = LTO4
Pool Type = Backup
Recycle = no # Bareos can automatically recycle Volumes
AutoPrune = yes # Prune expired volumes
Job Retention = 12 months
Volume Retention = 12 months
Volume Use Duration = 3 weeks
Next Pool = Offsite # consolidated jobs go to this pool
Storage = T-LTO4
}

Now because you only want a full once a week, and nothing more involved you would control that with a Schedule. In your case it would look like
Schedule {
Name = "ServerCycle"
Description = "For servers and other over night backups"
Run = Full sun at 1:05
#Run = Differential 2nd-5th sun at 23:05
#Run = Incremental Jan-Dec sun-sat at 2:05
}

You can use the same schedule with multiple jobs, given everything I’m doing here only runs one job at a time, all 3 servers will kick off at the same time, but 2 will wait till the first one finishes, next and next etc. So it’s ok to reuse the same schedule for multiple jobs.

Higly recomend you read the documentation closely, and’t don’t jump right into modifying to match what I pasted here, make sure you client/dir/storage can all talk to each other

status client
status dir
status storage

Once they can all talk to each other, then setup the storage make sure it shows everything correctly, then start modying filesets/jobs/schedules, lastly mess with volume settings, because that’s the thing I personally got hung up on the most often.



Brock Palen
bro...@mlds-networks.com
www.mlds-networks.com
Websites, Linux, Hosting, Joomla, Consulting

> --
> You received this message because you are subscribed to the Google Groups "bareos-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to bareos-users...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/bareos-users/49d17e62-dee7-4a3f-a86c-afb7a4188c45n%40googlegroups.com.

[Mi Zi]

hello plane ... big thx for your time to share this informations ... i will check this an comment my experiences

[Mi Zi]

hi Palen

sorry for the delay ... but better late than never :)

here is some comment with regards spooling feature:

this doesn´t work for me until i set the following parameter on Job level: see here Spool Data (Dir->Job)

Spool Data = yes

this setting can be overwritten in a Schedule: see here Run (Dir->Schedule)

the rest works like a charme at present time.

greetings michael

[Brock Palen]

Great glad it worked,

Yeah I forgot about that, I add that to my template and then I disable spooling in specific jobs eg archive jobs that I know the system is fast enough to saturate the tape drive.

Welcome to Bareos!

Brock Palen
bro...@mlds-networks.com
www.mlds-networks.com
Websites, Linux, Hosting, Joomla, Consulting

> To view this discussion on the web visit https://groups.google.com/d/msgid/bareos-users/e41789f0-88e2-41c3-8a09-84d4b6c39ff5n%40googlegroups.com.

[Mi Zi]

hi again,

and now the result of the performance with spooling:

it is SLOWER than without in my situation: (LTO-8 drive)

so i can recommend that you check both settings (with and without spool parameter)

with spooling parameter:
in my case i canceld the job after 11h (2TB stored)
without spooling parameter:
round about 9h for 3,7TB

greetings michael

[Spadajspadaj]

The point of spooling is to mainain constant rate of data for the storage device (even if it's done in bursts). It works best for jobs that have low effective transfer speed (for example - for incremental jobs over a huge filesystem where only a very small subset of files changes - the job checks every file and backs up only the changed ones so the job itself takes quite a long time but the effective data rate is very low since only a small subset of files is transferred). If you have a source which can easily fill your storage device bandwidth, you can skip spooling since it would effectively - as you observed - increase the backup time since the backup data would firstly be transferred into your intermediate spool directory and after that would get copied to tape - that's one data transfer too many.

To view this discussion on the web visit https://groups.google.com/d/msgid/bareos-users/898de9b4-afb9-4762-bcb7-3a7f643c9d7an%40googlegroups.com.