|
Re: [Bacula-users] Pool per client
2011-04-13 12:35:26
Pablo,
If you’re backing up to disk, you can have as many Jobs running simultaneously
as you want/need. The trick here is to define not only a unique Pool resource for
each Client, but unique Storage, Device and Media Type resources as well. Each
Client will have its own resources, so there will be no contention. Following
are the relevant pieces of my configuration, snipped for readability.
I admit that when I was setting this up it seemed a bit bulky,
but it works. I back up many slow Clients to disk; their jobs all begin and
run simultaneously.
Of course, if you can’t back up to disk first, this is useless
to you.
#### CLIENTS ####
Client {
Name = DORCAS
Address = DORCAS.URTH.COM
}
Client {
Name = SEVERIAN
Address = SEVERIAN.URTH.COM
}
#### JOBS ####
Job {
name = "DAILY:severian"
client = "SEVERIAN"
pool = "SEVERIAN"
}
Job {
name = "DAILY:dorcas"
client = "DORCAS"
pool = "DORCAS"
}
#### POOLS ####
Pool {
Name = "SEVERIAN"
Pool Type = Backup
Storage = "SAN:severian"
next pool = "LTO4"
LabelFormat = "severian-"
}
Pool {
Name = "DORCAS"
Pool Type = Backup
Storage = "SAN:dorcas"
next pool = "LTO4"
LabelFormat = "dorcas-"
}
#### STORAGE ####
Storage {
Name = "SAN:severian"
Device = "SAN:severian"
Media Type = "FILE:severian"
}
Storage {
Name = "SAN:dorcas"
Device = "SAN:dorcas"
Media Type = "FILE:dorcas"
}
And in bacula-sd.conf:
Device {
Name = "SAN:severian"
Media Type = "FILE:severian"
Archive Device = "/dp-SAN/severian"
}
Device {
Name = "SAN:dorcas"
Media Type = "FILE:dorcas"
Archive Device = "/dp-SAN/dorcas"
}
Greg
From: Pablo Marques
[mailto:pmarques AT miamilinux DOT net]
Sent: Wednesday, April 13, 2011 11:01 AM
To: Kurzawa, Greg
Cc: bacula-users AT lists.sourceforge DOT net
Subject: Re: [Bacula-users] Pool per client
Thanks Greg,
But I would still have the problem that I need a device tied up backing up each
client.
The problem I am facing is that I need to backup lots of slow clients, and I
need to come up with something so I can back them up all at the _same_ time on
one or maybe a few devices, and still have a Pool per client.
Pablo
From: "Greg Kurzawa" <GKurzawa AT pamida DOT com>
To: "Pablo Marques" <pmarques AT miamilinux DOT net>
Cc: bacula-users AT lists.sourceforge DOT net
Sent: Wednesday, April 13, 2011 9:29:31 AM
Subject: RE: [Bacula-users] Pool per client
Hi Pablo,
If you have enough disk space handy, you could send each
Client’s data to its own disk Pool with its own Next Pool specification.
Each Client’s data would be in its own Pool on disk, then move to its own Pool
on tape. This is exactly what I’ve done at my site, except the disk Pools
all point to the same tape Pool.
Greg
From: Pablo Marques [mailto:pmarques AT miamilinux DOT net]
Sent: Wednesday, April 13, 2011 7:28 AM
To: Randy Katz
Cc: bacula-users AT lists.sourceforge DOT net
Subject: Re: [Bacula-users] Pool per client
Thanks Randy,
Unfortunately Maximum Concurrent Jobs won't cut it because I need a different
tape per client. Still I set it to 1000.
When spooling is enabled, bacula still wants a tape from the client pool on the
drive before the spooling starts.
I could create a virtual library with MHVTL and use several drives or use a
disk-changer emulator. But I am not sure how scalable these solutions are.
Pablo
From: "Randy Katz" <rkatz AT simplicityhosting DOT com>
To: bacula-users AT lists.sourceforge DOT net
Sent: Wednesday, April 13, 2011 6:08:54 AM
Subject: Re: [Bacula-users] Pool per client
On 4/12/2011 4:04 PM, Pablo Marques wrote:
I have a setup to backup lots of
clients over slow links.
I want to have each client (or group of clients) backed up to dedicated client
pools, so client1 will go to pool client1, and so on.
That way I have better control of the space used, if a client goes away I can
simply delete the tapes (or files) an get the space back immediately.
Also it gives me better control on the retention on a per client basis.
The problem is that when I try to backup multiple clients at the same time, the
storage process has to wait for each job to finish before it can move to the
next because it needs to change the tape (different client --> different
pool). Some clients may take many hours to finish, forcing everybody else to
wait.
I enabled spooling, but it seems like Bacula requires to mount a tape from the
client pool on a drive before the client spooling can begin.
Can this be avoided?
A possible solution would be to do all backups on a special pool and after they
are done migrate later each client job to each client pool.
But I cannot find a way to modify the "Next Pool" dynamically.
It is a fixed setting on the Pool definition.
Does anybody have suggestions on how to accomplish this?
Look into Maximum Concurrent Jobs
in your Storage definition.
Regards,
Randy
------------------------------------------------------------------------------
Forrester Wave Report - Recovery time is now measured in hours and minutes
not days. Key insights are discussed in the 2010 Forrester Wave Report as
part of an in-depth evaluation of disaster recovery service providers.
Forrester found the best-in-class provider in terms of services and vision.
Read this report now! http://p.sf.net/sfu/ibm-webcastpromo
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|
------------------------------------------------------------------------------
Forrester Wave Report - Recovery time is now measured in hours and minutes
not days. Key insights are discussed in the 2010 Forrester Wave Report as
part of an in-depth evaluation of disaster recovery service providers.
Forrester found the best-in-class provider in terms of services and vision.
Read this report now! http://p.sf.net/sfu/ibm-webcastpromo _______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|
|
|
