I am trying out bacula and reading the manual, I have some questions
about what people do as best practice wrt to multiple clients and disc
volumes.
How big an issue are concurrent jobs being streamed to disc? Should I
most certainly always avoid this?
In 13.5, its suggested that each client
Is it possible to write includes in the dir conf file
to include any combination of resource types to keep the
single file from growing large and to segment configuration
from just a viewing/edit perspective?
Thanks,
jlc
You can use rsync to mirror the volume files. You can even schedule
this in a bacula job so that the synchronization happens after all
other jobs. See how the catalog backup works for the idea.
Ahh, priority being the ticket here. I will just shell script this then
with the catalogue backup.
The sigs for the gpg key on the sourceforge download don't
match the sigs on the el5 rpms, anyone know where to get the
proper key?
Thanks!
jlc
--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008
Anyone know of existing Bacula 3.0.2 packages for these
two distros?
Blastwaves are old and I don't want to compile from source
on these production machines.
Thanks,
jlc
--
Let Crystal Reports handle the reporting -
Is this not possible/wise? No matter what I do, following the dvd setup
my backup stalls waiting for me to mount, then label media? I would rather
have bacula mount the location when it needs to write to it.
Thanks!
jlc
Is this possible to do so that mysql binaries and libraries are not
needed?
Thanks,
jlc
--
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and
You need to better describe your problem. Post error messages and
describe what media you are using. I am confused with you talking
about both NFS and dvd.
John,
Sorry for the confusion, I am trying to setup a device in bacula-sd.conf
that behaves like a DVD but for NFS. I just hoped it would
The user community should make an effort to offer pre-compiled fd's and sd's
for various OSes. I was in the same dilemma.
Did you just manually compile the whole thing?
--
Let Crystal Reports handle the reporting - Free
Autofs should do the trick although I have never used it this way with bacula.
Yea, much simpler:)
Any reason why you do not have a bacula SD on the NFS server?
Because I am trying to avoid compiling on this Solaris server and I
can't find 3.0.2 packages in any Solaris repo:( I do want to get
Alas not, because building the SD also builds the tools like bscan, which do
need a database.
However, you can build Bacula with sqlite instead of mysql as long as you only
need the bacula-sd program. Sqlite can be built from source in the depkgs
download.
Thanks for that info! Is it feasible
After manually deleting the bacula mysql db, if I perform a
status storage=name in bconsole, I see old terminated jobs.
Where is this coming from?
Thanks!
jlc
--
Let Crystal Reports handle the reporting - Free Crystal
I always compile from source on production Solaris 9 10 machines.
I learned my lesson ages ago with doing such things on with
distros that utilize a package manager (even though Solaris 10's is
worth sh!t). From that day forward I try my hardest to always use the
package manager, and it looks
With most of my jobs, Bacula handles the Full on an off day if
required based on the previous status of the backup in the pool.
So I have one job, and the schedule determines when the Full/Diff
is done. I am trying to accomplish the same thing with my exchange
backup which uses a RunBefore script
Really? Where does it say that, exactly? I looked at what I believe in
the relevant section of the manual, and I came away with a completely
different understanding.
In the middle of general Functionality of
http://www.bacula.org/en/dev-manual/Variable_Expansion.html
But, now that I read it
Unfortunately, this is more or less irrelevant to what you're doing.
You need to be looking at the section of the manual that details the Job
resource, specifically the RunScript directive:
http://bacula.org/3.0.x-manuals/en/install/install/Configuring_Director.html#SECTION0063
I have a schedule defined as follows:
Schedule {
Name = Server Data Weekly Cycle
Run = Level=Full mon at 18:00
Run = Level=Differential tue-fri at 18:00
}
A job as follows:
Job {
Name = Backup-Data-client
Client = client-fd
JobDefs = Default
Pool = Svr_Data
Storage = name
When you ran the job the first and second time, do you mean you
initiated the job manually (i.e. with the run command in the console)
or do you mean you sat back, did nothing, and let bacula start the job
all by itself based on the schedule?
I used the run command (Well weBacula did actually). I
You're not actually testing your schedule settings when you run the job
manually.
Cedric/Mike,
Thanks for clearing all this up, that makes sense about not using the schedule
when
executed manually and the default behavior when not specified (As I was
specifying
it in the schedule).
It's now
I have MySQL installed and used a configure script like so:
CFLAGS=-g ./configure \
--prefix=/opt/bacula \
--sbindir=/opt/bacula/bin \
--sysconfdir=/opt/bacula/bin \
--with-mysql=/usr/mysql/5.1 \
--enable-smartalloc \
--with-pid-dir=/opt/bacula/bin/working \
Thanks for clearing all this up, that makes sense about not using the schedule
when
executed manually and the default behavior when not specified (As I was
specifying
it in the schedule).
I spoke to soon, looking at the schedule, it correctly showed the
intended Job Type. It started with a Full
I would suggest running in debug mode (-d).
The Init script changes to the correct user. So run the init script with
pfexec.
You should also check for the right permissions on the logfiles. Maybe it is a
problem with connecting to the database.
Tries the -d switch, I am only running the Storage
Did you test the config file?
/path/to/bacula-sd -t -c /path/to/bacula-sd.conf
Yup, no warning.
I run bacula-sd as user group bacula
Heh, turned out that the install scripts never created the
pid directory! Why there was no error, who knows:)
Working now!
Thanks everyone!
jlc
I spoke to soon, looking at the schedule, it correctly showed the
intended Job Type. It started with a Full on the specified time in
the Schedule and the next few scheduled times after that had specified
Differential's so the schedule showed this.
When it came time to run it (on its own as per
I am trying to deduce how to recover from just replicated disc volume
sets in the event the catalogue or entire Bacula server is lost.
What are the ramifications or requirements surrounding the placement of
the volumes on the new server? Can the Device I place them in have a
different Name and
I have several mixed clients all backing up to Linux and Solaris
SD's. The windows fileserver's are taking far too long, so in looking
at this, I noticed the cpu utilization on the Windows FD's isn't very
high but the SD are, why is that? Searching the forum showed compression
is done at the FD
To me everything looks good. The CentOS/el5 RPMs should be signed with my key
0xFAF24CCA which is available as
http://sourceforge.net/projects/bacula/files/rpms-contrib-fschwarz/rpmkey/fschwarz.asc/download
- Can you tell me which RPMs specifically had that problem?
- Which key was used to sign
cpu high on sd ? Did you use a soft raid ( level 5 ? )
or wrong params for the fs .
No raid, actually this was a test box setup with a kickstart file
using /tmp as the device location. Bizzare...
jlc
--
Let Crystal
Are dir also running on this machine ?
Nope.
So perharps the db config is too low ( default conf tend to be friendly with
hardware resources )
and if batch-insert is enable, bacula write the batch table also in /tmp so
there's concurrency on the same drive.
Reading up on this in the manual,
Is it possible to manipulate the fileset based on the job? For example, a
RunScript parameter has a %l to pass the Job Level on, can the fileset somehow
be manipulated like this as well?
Thanks,
jlc
--
Come build with
a RunScript parameter has a %l to pass the Job Level on
Hey, that sounds interesting, I did note eaven know that, can you point
me to the on line documentation where I could find more about that?
Hannes,
The area in the docs is under the Director Config Job Resource:
Any hints concerning the restore?
I very much encourage you to read MS whitepaper on Exchange DR.
Its backup software agnostic, but provides lots of info, its
invaluable!
--
Come build with us! The BlackBerryreg;
I am trying to execute a simple scp of the bootstrap files after the catalog
backup.
It always says:
ClientAfterJob: /var/lib/bacula/*.bsr: No such file or directory
When run as root at the shell, it works fine. As the director runs as bacula, I
checked perms, and /var/lib/bacula and all the .bsr
Hi,
Shouldn't it be RunAfterJob, since the bootstrap files are on the director?
On the JobDef, where are you putting the bootstraps? ( for me, in the default
JobDef there's 'Write bootstrap = /var/lib/bacula/%c_%t_%n.bsr ' - %c =
client name, %t = job type, %n = jobname )
Cheers,
Hmm,
I had
That’s because the shell interprets the * glob but bacula doesn’t.
The shell will replace /var/lib/bacula/*.bsr with a list of matching
file names (if these files exists) while bacula will send the string
as is to the program. The solution is to put your command into a
shell script and execute
I have a job that runs either fulls or diffs, sometimes when
there is a problem on the client, the diff will hang. That
being the case I figured I would config a max time for the diff.
Looking for a Differential Max Run Time like Incremental Max Run Time
but there is only a Differential Max Wait
There are now 8 volumes so what is the correct way to bring this back
to how I intended.
I should have mentioned that I did run Update/Volume parameters/All Volumes
from all Pools update/pools as well, so that leaves me with just removing
the volumes. I am guessing I have to see what's in the
I have mistakenly created a pool with Maximum Volume Files = 5
instead of Maximum Volumes= 5 as I meant. There are now 8 volumes
so what is the correct way to bring this back to how I intended.
Also, given that one of the jobs writes ~400,000 files, how did they
all end up in one volume with this
I hope you also changed the configuration file and reloaded it -
otherwise, newly created volumes will get the wrong settings again!
Yup, I restarted the director before running the updates in bconsole.
Obviously, waiting until the volumes are automatically recycled is the
easiest approach. If
Yet the default backup jobs always fails.
Well, what does the log say about why? Is it looking for a pool
that doesn't exist or have been fully defined?
I wish to backup to a device which is mounted on /backup - it's a NAS with a
good few TB of storage - how would I go about doing this? I am
Hi, I have a very simple question. Can bacula be used to back up to a NAS
instead of a tape drive?
http://wiki.bacula.org/doku.php?id=faq
#2 :)
It'll do just about anything,
http://www.bacula.org/3.0.x-manuals/en/install/install/Storage_Daemon_Configuratio.html#SECTION0083
Does anyone know if the reload command really works in bconsole?
For the manual, only works in some situations. Has anyone tested
this command?
From my highly limited and inexperienced use of it, it often failed
more than worked for me.
Currently when I run it, I get:
Cannot open config file
I have used this over 100 times without problem in the 5 years I have
used bacula. Well that is if I do a test first.
bacula-dir -t /etc/bacula/bacula-dir.conf
In the past if you had an error in your .conf file it would crash the
director so I always test first..
Well, I run the test w/o issue
What are the permissions of /etc/bacula?
Ugh, how could I have missed that :)
Fixed and working...
--
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend
We're thinking about buy a LTO Ultrium HP external drive.
FWIW, I have had very bad luck w/ HP branded LTO's, for several years
they had low MTBF and not last long for me...
--
Let Crystal Reports handle the reporting -
What do you use insted.. our former Quantum changer was equipped with HP
tape-drives, ditto is our current StorageTec SL500. From my limited view
HP is inside everything..
After the last library tanked, I migrated to Bacula and disked based backups
to 3 different servers, 2 of which are remote.
It's a Windows workstation and I've set up the system where the person can
just click a shortcut that activates his job and he gets the notification
about the job by e-mail later. So rsync is not an option.
Why not? I do a lot of bacula based backups from windows machines that utilize
cwrsync
I'm new to bacula and would appreciate some advice on securing mysql with
bacula. I have installed bacula-3.0.3 on Centos 5.4 with MySQL and all
seems to be working well. My only concern is how to add password
protection to the mysql database and not cause any of the bacula scripts
to stop
I would really appreciate it if someone could look over the bacula-dir.conf
file below which I took from the manual and modified. The manual example is
in chapter 25 'Automated Disk Backup' however when I used this bacula
returned an error about no default pool being defined. I therefore added
So it seems the bacula-dir.conf file is doing what it is supposed to do. Of
course I may be misinterpreting everything but it does seem to be
functioning as I need it to.
I found when I was learning it the most useful approach I had was
to actually just sit and read each Resource Definition in
I thought I would compile and test out 5 for a new setup I am migrating to.
As it's a test box, there is no password for bacula. It runs fine and the
pg_hba.conf has the modifications to allow access. I can run a catalog backup
as root, but bacula is failing with:
Can't get catalog information
From what I can see under the backups, VSS is successfully being used,
backing up files (open or closed), MSDE, IIS, and various other things
- just like I would like to expect.
Looks fine to me...
What I would like to know however, is how to restore these VSS snapshots?
What to understand
For such a full backup, you need some secondary windows installation
for the resore itself - for Server 2003 and XP, BartPE is a convenient
way to get such a thing. For newer Windows versions, you probably best
use Microsoft's PE system (I haven't actually built such a beast yet).
I cant say
I tend to disagree - but I admit you seem to know what you're talking
about :-)
I tend to disagree with that:) Heh...
Anyway, my scenario in more detail - I'd be happy to see any hidden
pitfalls!
Use a secondary windows /typically PE-based) to boot. Create the
partitions you originally had on
I am trying to take a mysql version 11 db and import it into a postgre db for
later
update to version 12. I exported as per the manual,
`mysqldump -u root -p -f -t -n bacula bacula_backup.dmp`
After manually creating the postgre db and verifying it, I attempted
`psql -Ubacula bacula
rpmbuild --rebuild --define build_centos5 1 --define build_mysql5 1
--define build_bat 1 --define build_wzconsole 1 bacula-5.0.0-1.src.rpm
was the command required to build my RPM's which installed without a hitch.
Because I was doing an upgrade, I had to run the update scripts, I didn't
I would like to be able to break out my bacula configuration into a more
'logical' (from my perspective) set of files. For example, all my clients,
I'd like to put into a file called bacula_clients.cfg and then reference that
file in my DIR config - same thing for my schedules, I'd like to do
Is there something way to do a Job backup run against more than one
client???
Example:
I have tree clients, and I wanna make just one Job backed up this tree
clients...
How can I performer this jobs??
Thanks for any idea...
Gilberto,
How many is many? Create a job def, then reference it, so
This is apparently a question on CentOS developers. Bacula project has
no powers into the matter.
As far as I know, an rpm's dependency is well, derived by he who writes
spec. devel for building, but server for install? Maybe the spec needs
fixing.
Hi all. I install bacula 5.0.1 and bat in Debian lenny. Director, storage and
client are in the same machine.
I execute /sbin/bacula/bacula start
but bacula no listen in 9101 ( yes listen in 9102 and 9103)
What can i do?
Well, you could check what's running, I don't know or use Debian, but
No,
mine are all 32bit
Look through your system logs, I am willing to bet there
is an error that explains this in there somewhere recent.
--
___
Bacula-users mailing list
Thoughts?
-Steve Polyack
You can attain this now in the meantime:
http://www.bacula.org/5.0.x-manuals/en/console/console/Bacula_Console.html#SECTION0028
That would be useful for a program which can have
several million lines of output from one command...
Although maybe a sql
The problem we've hit is the message ERROR: Cannot execute bconsole. File not
found. that shows up on the web browser. See attached image.
Since I'm new at this I don't know where to start looking and would appreciate
any pointers you may have.
There is a webacula users list.
Your problem
What about using osql.exe to backup your databases? It should work quite
seamlessly :)
That's exactly how I do it with a run before script, works perfectly.
--
___
Currently, it is impossible for me to build the rpm packages for CentOS5
x86_64, because you added readline-devel as a dependency into the spec file.
That package does not exist on CentOS5/RHEL.
I always thought bacula picks the readline lib during configure stage, so why
is this dependency in
I was wondering if anyone has installed bacula-fd via group policy. I
would be glad to know if anyone has already written a guide to using
bacula in association with ad/gpo for a deployment
Someone on the dev list just posted source for an opensource msi building
tool so that the resulting
I'm seeing this on a regular basis where the Windows installer will not create
the bacula-fd service and I'm having to do it manually.
I see it only on Win2k machines but as I only have a couple I resorted
to doing it manually and left it at that.
If you encounter a resolution, post back...
How would you restore a VSS snapshot without having VSS ( in you linux live cd
) ?
That's the real question.
So yes yours steps are naive (in my opinion). What is described in the wiki
are the rights step.
Since when do you need vss *after* the backup was created using vss? Vss simply
allows
We are dealing with issue that time to time periodic VSS snapshots produced by
Windows scheduler are disappearing after Bacula backup (FULL or INC or DIFF,
VSS enabled in the FileSet definition).
Any idea about this sporadic disappearing of Windows VSS snapshots?
Let me guess, you windows
sorry 9103
As far as SD's go, my experience is that they do not bind to a specific
address
unless told to, but then again, the case I refer to had two interfaces on the
same subnet. SDAddresses should still accomplish what you need...
My bacula server is multi-homed.
I am needing port 9102 to answer on all of the IP addresses on the server in
order to service all of the subnets attached to the machine
I looked at taking care of this on the network routing level but it just would
not be practical
What's funny is the
I have a fair number of failed jobs in my database, and I would like to
know is it possible to purge all the bad backups, so that all the
diskspace that has been taken up can be released?
from bconsole, `del jobid=#`.
If you have many, an sql query to look for JOB ID's with FAILED status piped
What would be a method of restoring a remote linux based server to it
last full backup, (or even it's first!). Is such a move possible?
Got a management card in it:)?
Seriously, there are a few ways all not without big chance. You can create
a default grub entry to boot into an instance that
Indeed, I'm using CentOS 5.5 amd64. I was not doing the installation
with RPM packages, but compile the source code obtained at the official
site of Bacula. On sunday, doing tests comparing with the installation
done with packages RPM, I realized that the difference was in which the
libraries
I manually did a Full on one job Tuesday and last night Bacula
claims there is no Full on record so it upgraded the Inc to Full.
I can see both jobs in different file vols and can restore files
from the first full?
Bacula 5.0.2 / CentOS 5 x64
Thanks!
jlc
Have you made any changes to the Fileset?
Phil,
Heh, it's been a busy last few days... I forgot we moved a slew of
data so I added an exclusion. Totally slipped my mind, the growth
in storage alert that was unexpected this morning caught me off
guard.
Thanks:)
jlc
I have a job scheduled some few hours from now, any way to modify a parameter
of it (diff-full) through bconsole w/o editing the conf files?
Thanks!
jlc
--
Automate Storage Tiering Simply
Optimize IT performance and
I have a W2008r2 server w/ 5.0.3x64 at has recently been nearly impossible to
backup. Its running
on HP hardware and I have tried updating the NIC firmware and drivers w/o
success. What's
getting more frustrating is that the 5.0.3 director on CentOSx64 takes forever
to cancel the job
when
sudo rpmbuild -bb --define build_postgresql=1 --define build_centos5 1 --
define rhel_version=501 bacula.spec
Not only do you not need to unpack the srpm to get the the spec to build from
But you shouldn't at all build as root?
Use a buidroot [1]:
$ rpmbuild --rebuild --define build_centos5=1
I believe that is a memory leak. Although I would not worry about 48 bytes..
John,
Thanks for the confirmation.
jlc
--
Start uncovering the many advantages of virtual appliances
and start using them to simplify
I needed to setup a longterm pool with another storage backend so I
created a new Storage resource in the Director conf with its own Device
and Media Type. The SD has a new Device to match, finally I created a
Pool resource and manually labeled the 4 file volumes I needed.
I attempted to run a
Post the output of the command. And
list media pool=PoolName
John,
Here is the Job Status:
2010-09-17 10:32:13 cluster01-dir JobId 282: Start Backup JobId 282,
Job=Backup-Data-Napoli.2010-09-17_10.32.11_03
2010-09-17 10:32:13 cluster01-dir JobId 282: Using Device FileStorage-LT
2010-09-17
That is the problem.
update pool from resource
and follow the prompts
update all volume in pool
again follow the prompts. I am going from memory here so the command
may be different..
John,
Something was awry with this catalogue, I don't know how that got set
incorrectly
so I just moved all
i'm trying to run bacula-fd client in windows 2008 r2 x64 and is not
working.
E program just don start, even the process show up in process list.
Anybody have any tips??
Firewall?
--
Start uncovering the many advantages
How can I reset the info in the database (or locally on the director)?
http://www.bacula.org/5.0.x-manuals/en/main/main/Installing_Configuring_MySQ.html#SECTION00423
--
Start uncovering the many
I have several physical and virtual fd's that are just unreliable to backup
against two sd/dir's all at 5.0.3 running CentOS5x64.
Anyone else having these issues? I get random network IO failures as
suggested by the dir?
Thanks,
jlc
I have seen this several times. Particularly with slow laptop clients
and compression turned on. Some NIC drivers appear to shutoff power to
the network interface after a period of network inactivity, and I
believe the inactivity is probably due to the laptop being too
underpowered to compress
May you help me with any suggestion?
Thanks in advance.
Follow my post of 59 minutes ago:)
Seriously, I am having the same issues and they are to intermittent on
my end to say for sure which level they appear on. Hit and miss...
Are you spooling? Try a 5GB spool file.
Doubt that's the issue, I am writing to a raid array w/ max jobs=1, not tapes,
so it would
be merely redundant for me and I am also having the issue.
I recall Kern suggesting the VSS code was very old, so I tried running for a
week w/o vss
and it didn't
Does this (network timeout) happen randomly or is it predictable?
Randomly for me,
Sometimes Fulls or Diffs fail at different points and sometimes they don't even
start transferring data, just sit in never-ever land until they timeout... Once
and
a while they screw with me and just finish:)
I'm curious if you have solved the issue or if you still have it. Also are
you running Symantec products?
Hey guys, keep the reply's on list for the benefit of the community. This
thread is a
prime example of other effected users.
Heavens no, I think even the proxy blocks Symantec, just to
Tell them to turn on TCP Keepalives on both sides (Linux its a
syssctl(8), WinNT by hacking random registry entries and setting them to
arbitrary and random values)
Up the TCP state expiration timeout values on any stateful L2/L3 devices
en route.
Not buying that, the default for Linux and
I suspect there is a more elegant way to accomplish this from a single
Bacula dir...?
Sure, several different schedules each with its own pool. Each unique
Schedule can have its own Pool directive.
jlc
--
Start
Anyone know the proper way to do this w/ an MySQL db?
I really don't know much about sql or bacula's db structure but
a ` DELETE FROM Client WHERE name LIKE '%host-fd%';` left the
db in an unstable state where a restore was needed. Volumes
could no longer be used that were expected to be available
I do a mysqldump from the old server's catalog and push that onto the new one,
copy the configs and change the volumes + pools, paths, etc accordingly?
Sounds good.
Will I run into problems because of the different versions?
Check the manual, there are update scripts for your db.
jlc
Both are linux (Centos 5.3)
Well I use CentOS and don't have this issue so I am guessing it is in fact
physical
issues you have...
--
The Next 800 Companies to Lead America's Growth: New Video Whitepaper
David G.
Is there a way to get each ServerName/Volume name to cycle its own set
of sequential numbers?
Should I care about this? Why or why not?
/snip
Maximum Volumes = 45
Recycle = yes
I can't comment on whether or not you*should* care, but if you do I see
a similar config to one specific setup
@Joseph, thanks, however I am a bit confused, I will end up with many
Pools, correct? Will they have different names? How does the Job select
each Pool?
Nope, you create the 45 volumes in each pool as per your config snippet.
As for my comment about whether I *should* care has to do with
If not, what is the correct procedure to do this within Windows?
How do you list the complete filesystem so that everything shows, are
there specialist tools for this? (it appears to be impossible to access
some dirs even if using console/administrator and explorer/DOS show a
different view
The bacula-fd daemon (according to netstat -na) doesn't appear to be
listening on the IPv6 address.
Force it to listen on whatever address/port you desire w/ FDAddresses =
http://www.bacula.org/5.0.x-manuals/en/main/main/Client_File_daemon_Configur.html
jlc
1 - 100 of 151 matches
Mail list logo