, with
native uncompressed capacity of 12.8TB and transfer rate of 472MB/s. No
word on what interface is going to be required to accomplish a sustained
transfer rate of almost half a gigabyte per second. Direct PCI-Express
connection...?)
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM
On 02/08/11 10:51, Alessandro Bono wrote:
I need only for disaster recovery so for this situation performance is
not so important, but I understand this performance level is required by
tecnologie
Could you use eSATA?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
On 02/08/11 11:19, Eric Bollengier wrote:
USB 3.0 is available and is about 4.8Gbit/s (600 MB/s)... :-)
http://hothardware.com/News/USB-30-and-SATA-6G-Performance-Preview/
For very limited, bleeding-edge values of available. ;)
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM
. This feature was BROKEN AND UNSAFE in
5.0.1. It is my recollection that it is fixed, and works properly, in
5.0.3... but as with all things, test first before you put it into
production, just to be sure.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala
On 02/06/11 09:20, Dan Langille wrote:
OMG
Pay no attention to the man behind the curtain.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free
be made to work properly in all cases. Use Full Pool,
Differential Pool, Incremental Pool directives in the Job or JobDefs.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix
is set absurdly short.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
Jobs have different retentions on their
Pools? If it is a BUG, is anyone working on it?
All I can say is that I have never personally observed this behavior.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p
, if they're *not* on the same
machine, the Catalog DB and storage daemon aren't competing for disk IOPS.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker
, consider RAID5. It will be slower, but you'll be
able to survive a single disk failure with a high probability[1] of no
data loss.
[1] Even RAID5/RAID6 has weaknesses. If you're interested, do a search
for RAID5 write hole.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
is the opinion of Bacula community about
the quality of LTO devices made by HP vs IBM?
Personally I've had better luck with IBM drives. In these post-Fiorina
days, my faith in the HP name is limited.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala
and install a separately-purchased drive of known
quality into it. A lot of off-the-shelf USB disk drives turn out to
contain the absolute cheapest possible disks, with predictable adverse
implications for both performance and reliability.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM
forward running to the new pools and let the old test
backups expire in their own time?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
just short of 24 hours or 7 days (or however long
you want to use a single Volume before moving on to the next one).
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin
to hold the maximum size of the number of jobs you might write
into it. A disk volume consumes only as much space as the data you
wrote into it.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
, you lose the ability to run
concurrent backups on multiple clients unless you use multiple Storage
devices.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl
On 01/27/11 07:33, Graham Keeling wrote:
On Thu, Jan 27, 2011 at 06:26:04AM -0500, Phil Stracchino wrote:
On 01/27/11 06:12, Graham Keeling wrote:
I think this last problem is what Phil is trying to solve by setting either
Maximum Volume Jobs or Volume Use Duration. But these solutions seem
your storage.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
volume, if available for creation.
If no new volumes can be created, then:
4. Eligible recycled or recyclable volumes, if any.
If none, then:
5. Ask the operator to resolve the problem by providing or creating an
appendable volume or purging an old one.
--
Phil Stracchino, CDK#2 DoD
with an
elephant gun.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
- is the invalid initial assumption that is screwing
you up.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years
. Bacula
is probably trying in vain to resolve that as a hostname.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's
on a single MySQL
instance is in practical terms limited only by how much hardware you
have under it. I support customers every day who have hundreds of DBs
(and sometimes, DBs with thousands of tables) on a single MySQL instance.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607
unpack and extract the requested file
on a local _fast_ disk and transfer that file over
network?
In short, yes.
More to the point, Bacula seeks directly to the requested file(s) on the
volume and extracts only them in the first place.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM
text compresses about 4:1 on average. Binaries, typically not
much at all. Digital images, digital video, zip archives, MP3 audio?
Forget it. Databases? Sometimes you get lucky and get 10%-15%
compression, sometimes they don't compress at all.
--
Phil Stracchino, CDK#2 DoD#299792458
On 01/15/11 01:43, Silver Salonen wrote:
On Friday 14 January 2011 23:30:15 Phil Stracchino wrote:
Stan:
Without commenting further on the reasonableness of your data owner's
ideas about how to do this, the real-world chances of the Bacula
director encountering an error increase with every
On 01/15/11 13:09, Radosław Korzeniewski wrote:
2011/1/15 Phil Stracchino ala...@metrocast.net
mailto:ala...@metrocast.net
No. I don't know enough about the internals of the Director to know why
it is unsafe, but experience has shown that it is. Sooner or later, a
reload
of running this backup as 12,000 individual
single-file backup jobs is still going to just about kill you, though, IMHO.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin
successive reloads without a problem is very small. The
chance of it surviving 12,000 is essentially zero.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl
)
You can reload while the Director is running. However, it is not 100%
safe. My advice: Try not to do it any more often than you have to, and
if you need to do a running reload, restart the Director at the next
good opportunity. If necessary, make an opportunity.
--
Phil Stracchino, CDK#2
On 01/10/11 02:49, Marcin Krol wrote:
Out of interest, which filesystem are you using on this volume?
xfs, as always when dealing with lots of small files.
But ... xfs is optimized for long streaming reads and writes.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607
On 01/10/11 01:56, Silver Salonen wrote:
On Monday 10 January 2011 03:14:22 Phil Stracchino wrote:
But then there isn't an SD named babylon4-sd2 to connect to.
What's the correct way to configure this?
The name of the Storage resource in Director configuration does not matter to
SD, so
On 01/10/11 09:40, Marcin Krol wrote:
But ... xfs is optimized for long streaming reads and writes.
The only reliable alternative for me is ext3 and its performance is
unacceptable when handling directories with lots of small files.
Have you tried jfs?
--
Phil Stracchino, CDK#2
is also
excluded from backup.
Yes, that's what he wanted: Any directory under SVN version control.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free
and then
the subsequent Differential Backups later, on top of the Full Restore,
but I don't know if this is the best solution.
Any ideas?
Sounds to me like you need your TPA and TPB volumes to be in different
Pools assigned to different storage devices.
--
Phil Stracchino, CDK#2 DoD
for each one of the Diffential
and Full backups...
That sounds correct. Do the Pool resources specify the correct Storage
devices?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
be using the Full Pool, Differential Pool,
Incremental Pool directives in your Job or JobDefs resources.
This won't affect the problem you're seeing, though.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p
or snapshot
that typically starts most DB backup schemes should be an atomic
operation that occurs with the tables locked, if another job starts
while your catalog job is running it really shouldn't matter. It's not
going to be able to modify the catalog while the catalog is being dumped.
--
Phil
). Which is just as well, since I
don't like the idea of relying on database or table locks.
Yes, disabling mixed priorities would prevent any higher-priority job
from starting while the catalog update was running.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala
Address = babylon4.babcom.com
Maximum Concurrent Jobs = 20
SDPort = 9103
Password = Some Password Here
Device = FileStorage2
Media Type = File
}
But then there isn't an SD named babylon4-sd2 to connect to.
What's the correct way to configure this?
--
Phil Stracchino, CDK#2 DoD
it, and you might want to consider submitting it as a
contributed support script.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
(thus making
certain there is a minimum of five minutes for attribute metadata from
the job to be flushed).
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl
everyone's gone home and taken
their laptops with them. All of the VirtualFulls would fail.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
to be created?
No, it should create no new media, as no new data is copied, only new DB
records.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free
.
$ .../bacula/etc/bacula stop
$ .../bacula/etc/drop-mysql-tables
$ .../bacula/etc/make-mysql-tables
$ .../bacula/etc/bacula start
Done.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man
about 230GB on a 200/400GB LTO2 tape.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
data such as digital audio
and video.)
The way LTO compression works, the tape drive compresses each block on
the fly, then writes to tape whichever is the *smaller* of the raw or
the compressed block, flagging it accordingly.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607
each backed-up file as a File record
(in the File table) which contains pointers to a Filename record and a
Path record. The complete pathname does not exist *as such* at any
single location in the database.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala
On 01/05/11 17:31, Arunav Mandal wrote:
Is deduplication possible in Bacula.
Not yet. :)
(Well, OK, not deduplication as such. Yet. But look into the Base Jobs
feature. It *MAY* do what you want.)
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala
in memory
during start up.
Right you are.
Ah, upon re-reading, I see I misunderstood the question
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl
time as you have Storage devices. Is
that really what you want?
This question comes up fairly frequently on the list. Most of the time,
people who want to do this want to do it for all the wrong reasons and
don't fully understand the implications.
--
Phil Stracchino, CDK#2 DoD#299792458
it? The first step is to get your drive
working at the OS level. If you can't do that, you'll get nowhere.
Does your kernel have SCSI tape support either compiled in or available
as a compiled module?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net
# Not compatible with 'Always Open=no'
AlwaysOpen = yes
RemovableMedia = yes
RandomAccess = no
Autochanger = no
Maximum File Size = 5GB
Backward Space Record = yes
}
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net
.
Bertrand,
Could you possibly be a little more specific than a bootstrap issue?
The actual failure message would be useful in diagnosing the problem.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
additional brokenness
of their own.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
a bootstrap file is
actually being created?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
is to be used to authenticate between
them.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's
.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
the modification date on all the files to
now. On any Unix-like system, 'touch' can do this for you. I'm not
sure how you'd accomplish it on Windows.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
at a time per storage device, because no storage device can have
more than one Volume mounted at a time.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker
be having to create new volumes and switch volumes
every few seconds.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's
Yearly disk pool (which you do), and you run that
job once a year (which one assumes you plan to), then it's going to be
about 200 years before Bacula has to purge a volume from it.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala
then.
Ah, I didn't realize you were using so many volumes per job. If you're
running multi-terabyte backup jobs, aren't 5GB volumes a bit small?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
}:${Minute:p/2/0/r}
RecyclePool = Scratch
}
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
On 11/10/10 12:01, Rodrigo N. de Castro Fernandes wrote:
Hi Phil Stracchino,
many thanks for your prompt answer.
Ok, i got it (the tape is automatically unmount when it fills). Is there
any way to run a script shell to eject the tape (mt -f /dev/st0 eject)
when it fills?
Why use a shell
On 11/08/10 18:32, Dan Langille wrote:
On 11/8/2010 5:18 PM, ganiuszka wrote:
Phil Stracchino pisze:
The DVD writing code in bacula proved to be both unreliable and
unmaintainable, has been deprecated in the last several major releases,
and is planned to be removed shortly.
Hi,
Where did
, or
8.5GB for dual-layer) disk volumes, then burn the volumes to DVD as a
separate operation using external CD burning software.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix
.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
? Perhaps the major
difficulty would end up being convincing linux to ignore the physical
disk-swap event, but that might depend on how the external enclosure
is set up and configured?
...So what you have to ask yourself is, do you feel lucky?
--
Phil Stracchino, CDK#2 DoD#299792458
On 11/05/10 16:49, Bruno Friedmann wrote:
On 11/05/2010 04:18 PM, Phil Stracchino wrote:
I think the real key thing to take away from this is, when you are
talking datasets of this size, YOU MUST HAVE REDUNDANCY in your storage.
Otherwise it's like playing Russian roulette with a revolver
and video originally in mind, but just the thing
for Bacula volumes. You might want to consider btrfs when it becomes
production-ready, though.
My backup server runs Solaris 10 x86, and backs up to ZFS.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala
?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
you're still new
to Bacula.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's the mileage
/BackupFC13
Media Type = File1
LabelMedia = yes
Random Access = yes
AutomaticMount = yes
RemovableMedia = no
AlwaysOpen = no
}
Well, I certainly don't see any discrepancies there. You did restart
ALL Bacula daemons, right...?
--
Phil Stracchino, CDK#2 DoD#299792458
= DiskPool2
Storage = DiskStore
Pool Type = Backup
# create only volumes of MediaType File2 in this pool
}
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin
-manual/Backup_Strategies.html, the daily tape
rotation. If I add the Volume Use Retention to 23h in this model, will it
keep using the same tape or will it actually rotate between the two tapes?
I think you mean volume use *duration* there.
--
Phil Stracchino, CDK#2 DoD#299792458
of your Catalog and storage
pools over time.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's
as:
SELECT DISTINCT Path, Name FROM File LEFT JOIN Path ON File.PathId =
Path.Pathid LEFT JOIN Filename ON File.FilenameId = Filename.FilenameId
WHERE Name LIKE %partial filename% AND Path LIKE %partial path%
ORDER BY Path, Name;
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
On 10/22/10 14:07, mark.berg...@uphs.upenn.edu wrote:
The pithy ruminations from Phil Stracchino ala...@metrocast.net on Re:
[Bacula-users] Searching for files were:
= On 10/22/10 02:47, ladolf wrote:
= Maybe anyone knows how to query the bacula database directly to get
= search result
, what I'm driving at
here is, do you have any capability (Linux LVM, for instance) to take a
filesystem snapshot, then mount a clone of the snapshot somewhere and
back that up?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p
and start again the services.
Are you certain that ALL Bacula processes were halted? None of them
were hung?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl
to recall discussion of this issue
in the past.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years, it's
a dump file at all, and will in fact
simply take filesystem snapshots for my incremental database backups and
keep the most recent snapshot around until the next night's backup has
been completed.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala
On 10/18/10 08:08, Damian Ge;bicki wrote:
Phil Stracchino wrote:
On 10/18/10 07:21, Holikar, Sachin (ext) wrote:
Well, without seeing the script, we're somewhat guessing. Personally, I
keep my catalog dumps around because if something crashes my database,
it's faster to just reload the last
to access every job since the last
full.
...Or the last differential and every incremental since the last
differential.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix
?
Thanks for any suggestions,
Sorry Simone, but once you've partially overwritten the tape, you're
pretty much out of luck with any out-of-the-box tools. You'd have to
write a specific program to scan the entire tape beyond the EOT mark to
try to recover any data from it.
--
Phil Stracchino
backup, it will not upgrade any *existing* backup.
However, if you delete *the most recent* Full backup, which has current
Incrementals or Differentials based upon it, then the *next* Incremental
or Differential that you run will be upgraded to Full.
--
Phil Stracchino, CDK#2 DoD#299792458
, if your backup load
is saturating your network, the only real way out of the problem is to
increase the speed of your network.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man
. The most likely reason is a
configuration error on the drive. What capabilities does mt/mtx report
for your drive?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin
On 10/11/10 16:48, Andrea Venturoli wrote:
On 10/11/10 18:32, Phil Stracchino wrote:
Can 4 hours and a half be considered a normal time to move to the end
of data in order to append to a volume?
Is there anything I should check?
That does not seem remotely normal. The most likely reason
is interested ...
Please do. This is interesting information.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
It's not the years
,
This is an interesting observation. How does one determine/set the
InnoDB block size?
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p...@co.ordinate.org
Renaissance Man, Unix ronin, Perl hacker, Free Stater
On 10/08/10 17:49, Attila Fülöp wrote:
please see
http://dev.mysql.com/tech-resources/articles/mysql-zfs.html#Set_the_ZFS_Recordsize_to_match_the_block_size
16K is the zfs recodesize I'm using.
Aha! Thanks, Attila. Exactly what I needed.
--
Phil Stracchino, CDK#2 DoD#299792458
On 10/07/10 04:10, Ralf Gross wrote:
Phil Stracchino schrieb:
On 10/06/10 14:35, Mingus Dew wrote:
John,
I think I had to create a bogus schedule, that bacula wouldn't
accept the job config without a schedule. I think I'll disable the job
in bconsole and try to start it remotely. Just
,
This is a solved problem. The following two existing directives added
to your JobDefs or Job resource will take care of this for you:
Allow Duplicate Jobs = no
Cancel Queued Duplicates = yes
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala
of using a larger, not smaller,
block size for applications such as this where you expect the major
usage to be extended streaming reads and writes.
That said, my own disk SD runs on ZFS with default block size and works
just fine.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607
, bacula version is 5.0.3 on CentOS
(compiled from source).
Uwe, if I run eight concurrent jobs on an LTO2, I'm sure you can run two
on an LTO4 without any problems.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala...@metrocast.net p
On 10/06/10 05:18, Robert Oschwald wrote:
hi,
I got a problem with automatic tape status change, which does not work. Tapes
always got status Append, even when their retention period is exceeded.
Have you tried setting a volume use interval on the tapes?
--
Phil Stracchino, CDK#2
of the correct architecture and OS
that has the database drivers, then install all the SD machines from
that build. Sure, tools like bscan that are linked against database
client libraries won't run; but if you're not going to try to use them
anyway, you don't really care about that.
--
Phil Stracchino
, you might create a new empty
Catalog, then simply use bscan to scan the volume containing the most
recent catalog backup into the new Catalog, then run a full Catalog
restore from that.
--
Phil Stracchino, CDK#2 DoD#299792458 ICBM: 43.5607, -71.355
ala...@caerllewys.net ala
701 - 800 of 1318 matches
Mail list logo