On Sun, 18 Jun 2006, Frank Sweetser wrote:
> On Mon, Jun 19, 2006 at 11:43:22AM +1200, Norbert Murzsa wrote:
>> After a successful building I have one running process on RHEL3 and
>> RHEL4 but I have three running processes on RH 7.3.
>
> They're not actually three seperate processes - they're thr
On Tue, 20 Jun 2006, Kern Sibbald wrote:
>>> Well, it really should not have much trouble backing them up or restoring
>>> them, though the restore may be a bit slower when creating so many files
>>> in one directory -- this is really OS and memory size dependent.
>>
>> It's the restore tree build
On Thu, 22 Jun 2006, Christoff Buch wrote:
> Might I add a few lines I've already added to Rumkos original mail
> before?:
>
> Spooling is very good, but shouldn't slow down backup since this affects
> per-night-capacity.
>
> Also, I thought about simultaneous jobs going on several drives of one
>
On Fri, 23 Jun 2006, Arno Lehmann wrote:
> This morning, I had a very unreliable WLAN connectivity (by now there
> are about 6 to 10 WLANs overlapping. Crap.)
Time to move to 802.11a, or start trying to sort out a local frequency
reuse plan (only channels 1,7,11 are non-overlappping at 2.4GHz a
On Sat, 24 Jun 2006, Christoff Buch wrote:
>> Bacuka has the ability to already run simultaneous jobs on different tape
>> drives and spool them to disk.
>> Despooling is triggered by the spool file reaching maximum size, or the
>> spool disk area exceeding maximum size.
>
> => What exactly do you
On Sun, 25 Jun 2006, Brian A. Seklecki wrote:
>> you do a "reload" command or whenever there is a sudden clock shift, which
>
> Do you postulate that large clock shifts on NTP synchronized systems are
> an indication of a low-level hardware problem?
It is common for initial NTP or rdate set failu
On Thu, 29 Jun 2006, Kern Sibbald wrote:
> I have released Bacula version 1.38.11 to Source Forge. This is a bug fix
> release, and most importantly, it has the scheduler patch integrated into it.
Kern (and others)
Will there be a RHEL 4 release? (or will FC4 be sufficient?)
AB
Using Tomcat
On Sun, 2 Jul 2006, Kern Sibbald wrote:
> regression tests, requiring a power off. When it came backup, my boot
> partition was empty. I loaded the latest kernel, made 3 full backups, and
> rebooted, at which time, it died a hundred horrible deaths (horrible death
> = a line of red output on the
On Mon, 3 Jul 2006, Kern Sibbald wrote:
>> Perhaps your hardware is slightly flaky?
>
> That is possible, but after having problems on 3 different machines that I
> upgraded from FC4 to FC5, I suspect that the problem is in their upgrade
> procedures that leaves the system in an unstable state.
W
On Mon, 3 Jul 2006, Kern Sibbald wrote:
>> As with your FC problems, I suspect my suse ones are due to continual
>> upgrading since 8.2 - a clean installation on a Mac Powerbook is running
>> more than happily.
>
> Quite possibly this was the problem with the FC4 to FC5 update, and I
> suspected t
On Wed, 5 Jul 2006, Arno Lehmann wrote:
> I'd recommend to try the mtx-changer script alone first. I suspect there
> is a timing problem in there. My experience is that it's usually quite
> useful to use the wait_for_drive function instead of a pre-determined
> timeout, because - especially with D
On Thu, 6 Jul 2006, Arno Lehmann wrote:
(re mtx-changer)
>> Is the changer script still using a preferred timeout by default??
>
> The current cvs version, at least, uses a default timeout of 15 seconds
> and _not_ wait_for_drive.
Kern, is there any compelling reason for not making wait_for_driv
On Fri, 7 Jul 2006, Kern Sibbald wrote:
>> Kern, is there any compelling reason for not making wait_for_drive the
>> default?
>
> The problem is that it is "system" dependent, and so on some systems would
> likely cause it to go into an infinite loop. If someone really wanted to
> work on it, and
On Mon, 10 Jul 2006, Christoff Buch wrote:
> Ok. But thanks for your time all the same!
>
>
> "user100" <[EMAIL PROTECTED]> wrote on 10.07.2006 16:10:37:
>
>> Sorry forget it: "Fatal error: job.c: 1617 Bad response to Append Data
>> command. Wanted 3000 OK data" - does not seems like the "permissi
On Tue, 11 Jul 2006, Kern Sibbald wrote:
> Your emails to any Bacula list will be dropped if they do not contain at least
> one of the following content types:
I believe they will also be dropped if the Message-ID format is invalid -
there must be a <[EMAIL PROTECTED]>
Many broken mail clients
On Tue, 11 Jul 2006, Christoff Buch wrote:
> Hi Alan
>
> Is 1.38.9 meant to be an older version?
Yes, current is 1.38.11
> 07-Jul 23:00 heem-sd: 3304 Issuing autochanger "load slot 15, drive 0"
> command.
> 07-Jul 23:00 heem-sd:
> A-1_dali-AAW-Con-Fin-Man_.2006-07-07_23.00.00 Fatal error
On Wed, 12 Jul 2006, Christoff Buch wrote:
> => Everything runs as root here... Don't even have an user "bacula".
That eliminates one issue.
>> If this works, then use mtx-changer and perform the tests detailed in
> the
>> bacula manual.
>
> => If you are talking about btape tests for drive / au
Nothing major here.
Kern:
When doing a 'status storage' it would be extremely handy to know
whether jobs are:
1: spooling to disk
2: flushing the spool
3: waiting to flush
This is only applicable when spooling is enabled (of course).
How hard would this be to implement?
I had 20 jobs run
On Fri, 14 Jul 2006, Julien Cigar wrote:
> Seems that I don't have such device ... :
>
> phoenix:/home/jcigar# tapeinfo -f /dev/sg0
> cannot open SCSI device '/dev/sg0' - No such file or directory
>
> I've always used /dev/st0 or /dev/nst0, but I also have the following :
/dev/st* doesn necessari
On Fri, 14 Jul 2006, Gabriele Bulfon wrote:
> Hello,
> I have some bacula installations on SunFire 280R Sparc machines, with Solaris
> 10.
> These machines apperar to be very very slow with respect to other
> installations (such as v20z) with same LTO2 device.
> As you can see from the report, 6
On Fri, 14 Jul 2006, Martin J. Green wrote:
> I labelled volumes using bconsole 'label' command and it appeared to
> work ok (and the files exist), however status still shows /mnt/storage
> as being non-existent...
It will until there's a job using it.
-
When running concurrent jobs and data spooling, there is no indication of
which jobs are being spooled or despooled
As an example, take part of the output of several adjacent jobs:
15-Jul 18:49 msslay-sd: User specified spool size reached.
15-Jul 18:49 msslay-sd: Writing spooled data to Volume.
On Mon, 17 Jul 2006, Gabriele Bulfon wrote:
> Thanks,
> this is very interesting.
> My LTO2 drives (I have many installed) are from Certance.
Mine are HP drives installed in a HP MSL6000 library (aka NEO4000)
> Do you achieve these rates on a SunFire 280R?
No, Wintel hardware (HP Proliant DL580
I'm not sure about this yet
It appears that when running spooling and concurrent jobs on an
autochanger with multiple tape drives, that the tape drives are being
locked on a per-changer basis and not on a per-drive one.
IE: Full spool files are only being flushed to one drive at a time, e
or jobs waiting to reserve a drive.
The same hangs occur in the tray monitor.
Having observed it for a while, it looks like anything using drive 1
causes drive 0 to be locked out, but using drive 0 does not lockout drive
1.
Kern/others: Any ideas?
> On 18.07.2006, at 16:04, Alan Brown
On Tue, 18 Jul 2006, Dan Horne wrote:
>> What happens if you define an autochanger resource and make
>> the two drives part of an autochanger? Even though there is
>> no _true_ autochanger, it seems to me that the logic would be
>> there to make it do what you want.
>>
>
> I don't see how this wo
0 # 10Gb
WritePartCommand = "/etc/bacula/dvd-handler %a write %e %v"
FreeSpaceCommand = "/etc/bacula/dvd-handler %a free"
}
Messages {
Name = Standard
director = Director-dir = all
}
>
>>
>>> On 18.07.2006, at 16:04, Alan Brown wrote:
On Wed, 19 Jul 2006, Mathew Brown wrote:
> Hi,
> I couldn't find much information regarding LTO drives and support
> besides the basic information provided by:
>
> http://www.bacula.org/dev-manual/Supported_Tape_Drives.html
>
> which shows that Bacula should work fine with LTO drives. However,
On Wed, 19 Jul 2006, Kern Sibbald wrote:
> Can you clarify one point? When you say that when the job is using drive0 the
> SD is fine, do you mean that it can use drive1 as well, but when using
> drive1, drive 0 cannot be accessed? I would be *very* surprised if the SD
> does not handle drives i
On Wed, 19 Jul 2006, Tracy R Reed wrote:
> Fortunately I stopped it before too much was deleted. But now I don't
> know exactly what files were deleted. Looks like it deletes in
> alphabetical order and I still have files named c* so I guess I only
> need to restore files [a-cA-C] and not director
On Wed, 19 Jul 2006, MaxxAtWork wrote:
> I have recently received an HP LTO2 (Ultrium-448) with 1x8 Autoloader,
> and everything seems to be working just ok, including the autochanger.
> My server is a HP DL380 G3 (with a SCSI HBA type LSI22320-R, taken
> from an unused Sun server...), system is r
On Wed, 19 Jul 2006, Kern Sibbald wrote:
> In looking over what you wrote and your conf file, I have the following
> comments:
> 1. It might be simpler if you pointed everything at /var/bacula/spool unless
> you are mounting the spooling subdirectories on separate filesystems. Bacula
> automatica
On Wed, 19 Jul 2006, Peter Eriksson wrote:
>> This seems to clear when the file being flushed to tape finally finishes,
>> but of course when running concurrent jobs there's always another file
>> ready to flush - if that is to drive0 then things are normal again (and
>> drive1 flushes happily too
On Thu, 20 Jul 2006, Kern Sibbald wrote:
> Large and growing backup data is a typical symptom of users who backup to disk
> and forget to exclude their disk Volumes.
Or users who are using spooling, but forget to exclude the spool area. :)
Been there, done that... (It's a bad idea to try and bac
On Thu, 20 Jul 2006, Luis Marcelo Achite wrote:
> 3910 Unable to open device "DVD_Writer" (/dev/hda): ERR=There is no
> valid media in the device "DVD_Writer" (/dev/hda).
>
> I can use dvd+rw-tools to view DVD features, also can fast blank a dvd
> using dd command, and can make a manual backup usi
This has been discussed in the past.
Basically., I need to switch off spooling for specific backup levels
(full) while continuing to maintain it for others (differential,
incrementals)
Someone posted a bacula-dir.conf Job directive which did that, but it
isn't showing in the manual.
Can anyo
On Sat, 22 Jul 2006, Tracy R Reed wrote:
> Alan Brown wrote:
>> There is a restore option to not overwrite existing files.
>
> But that would also restore a lot of deleted mail into my Maildirs which
> I would like to avoid.
Your choices then come down to:
1: using wxconsole t
On Sun, 23 Jul 2006, Shawn Robert Lesniak wrote:
> Greetings,
> I am the system administrator at a lab at UCR. We have a ~800GB RAID
> 5 array that was being backed up using bacula that suffered severe data
> loss (mostly due to reiserfsck) that was being backed up to an Exabyte
> tape changer
On Thu, 20 Jul 2006, John wrote:
> Everything works fine on the linux boxes but the solaris boxes are showing
> these
> errors:
These are normal. Look for the onefs keyword in the manual.
-
Take Surveys. Earn Cash. Influen
Item : Tray monitor window cleanups
Origin: Alan Brown [EMAIL PROTECTED]
Date: 24 July 2006
Status:
What: Resizeable and scrollable windows in the tray monitor.
Why:With multiple clients, or with many jobs running, the displayed
window often ends up larger
Item : Clustered file-daemons
Origin: Alan Brown [EMAIL PROTECTED]
Date: 24 July 2006
Status:
What: A "virtual" filedaemon, which is actually a cluster of real ones.
Why:In the case of clustered filesystems (SAN setups, GFS, or OCFS2, etc)
multipl
On Mon, 24 Jul 2006 [EMAIL PROTECTED] wrote:
> Is there a way to manually execute the algorithm that bacula uses to request a
> tape for recycling, in order to predict which tapes will be recycled next?
I wish.
I had 20 purged tapes (out of 40 recent purges) loaded in the autochanger
last night
1: What is the current intended behaviour of "update slots"?
I thought it was intended to scan slots and drives without requiring a
tape drive unload in 1.38.
Currently it does - which means that changing out tape sets cannot be done
unless at least one drive is idle/blocked - problematic in s
On Tue, 25 Jul 2006, Kern Sibbald wrote:
>> I thought it was intended to scan slots and drives without requiring a
>> tape drive unload in 1.38.
>
> I was going to remove the requirement to unload the drive, but the users
> objected to that so it currently unloads the drive you specify.
OK.
>> C
On Wed, 26 Jul 2006, Birger Blixt wrote:
>> Here are the tricky parts (Only 45 slots reported today - one magazine is
>> out)
>>
>>Storage Changer /dev/sg16:2 Drives, 45 Slots ( 2 Import/Export )
>> Data Transfer Element 0:Full (Storage Element 1 Loaded):VolumeTag = ALOW0026
>> Data Transfer
On Tue, 25 Jul 2006, Jo Rhett wrote:
> Okay, now it's been what - 3 weeks? And as far as I can tell nobody
> has even looked at this problem.
Bacula is user-supported software.
>From what I've seen on the list, you have provided inadequate information
about what is going wrong, have not spent
On Thu, 27 Jul 2006, Birger Blixt wrote:
>> When a tape is loaded directly into a drive from a mailslot, some
>> changers classify it as "unknown"
>> Data Transfer Element 0:Full (Unknown Storage Element
>> Loaded):VolumeTag = ALOW0041
>>^^^
On Thu, 27 Jul 2006, Emery Guevremont wrote:
>> At least, spooling should guarantee that your tape drive is able to run at
>> full speed continuosly. If for some reason your system isn't able to supply
>> data to the tape drive at fast enough rate, the drive keeps
>> stopping-repositioning-writing
On Tue, 25 Jul 2006, Jo Rhett wrote:
> This has to rank as one of the stupidest spam detection measures I
> have ever seen.
Their servers, their rules.
Deal with it.
-
Take Surveys. Earn Cash. Influence the Future of IT
Jo
On Fri, 28 Jul 2006, Kern Sibbald wrote:
> There are certain limitations in version 1.38.x, that are explained in the
> release notes I think -- mainly, if a tape is in drive 2 and Bacula is
> writing on drive 1 and then wants to write on the tape that is in drive 2, it
> will unload it from drive
On Fri, 28 Jul 2006, Jo Rhett wrote:
>> Their servers, their rules.
>>
>> Deal with it.
>
> Frankly, I can't. Having to use a GUI mail client makes participation in the
> 100+ mailing lists that I contribute code and provide assistance on
> completely untenable.
That's funny, Pine works perfe
On Fri, 28 Jul 2006, Kern Sibbald wrote:
>> One of the more annoying problems is that it will insist on recycling
>> tapes which are not in the changer, then block on loading them, when there
>> are purged recycling-candidate tapes available in the changer.
>
> Yes, this is probably to be expected
On Sat, 29 Jul 2006, Mark Nienberg wrote:
> I can see this coming up also when I decide a tape is too old and should
> be replaced.
> I'll need to substitute a blank tape and label it with a name that
> already exists.
IMO It's bad practice to reuse a tape label in that sort of instance.
Assu
On Sun, 30 Jul 2006, LDB wrote:
> I recently moved my backup server to another server. I copied
> over the DB like I was suppose to and the necessary bacula files.
> I also recently decommissioned a server and I have its old backup
> as well but I cannot restore it even though its volume exists.
On Mon, 31 Jul 2006, LDB wrote:
>> Assuming the server still exists in bacula-dir.conf, perhaps the file
>> tree has expired and been prined off?
>
> Yes, it does exist in bacula-dir.conf. How do I then get the file tree
> back?
bscan - which is time-consuming.
AB
-
On Wed, 2 Aug 2006, Kern Sibbald wrote:
>> Kern, is this fixed in a newer version?
>
> I believe more information is printed in 1.39.x, but there are so many changes
> that I cannot remember the exact case that was enhanced.
>
>> Can we get it fixed?
>
> If it isn't already fixed, then it will be
On Wed, 2 Aug 2006, R.I. Pienaar wrote:
> it's not optimal
Indeed: Multiple concurrent jobs on the same FD are still not easily
distinguishable.
> but you can get each FD to write a log on its own box
> by adding an append directive to its messages settings.
--
On Wed, 2 Aug 2006, Kern Sibbald wrote:
>> Do you know wether the clone method would run the same backup twice,
>> doubling both network and server load, or would the copy be cloned
>> locally on the backup server?
>
> It runs two separate jobs, so the backups will be done at the same time each
>
On Wed, 2 Aug 2006, Dirk Kleinhesselink wrote:
> On Sat, 29 Jul 2006, Kern Sibbald wrote:
>
> This solution solved the problem, thanks very much: I defined the
> media type for the tape drive in one of the autoloaders to be LTO-4,
> leaving the other one as LTO-3 and now I can restore from backup
On Wed, 2 Aug 2006, Dirk Kleinhesselink wrote:
> On Wed, 2 Aug 2006, Alan Brown wrote:
>
> Well, there is NO such thing as an LTO-4 drive.
Yet
It's in the LTO development roadmap.
> They are both LTO-3 drives.
> I'm not sure that media type declaration does anyt
On Wed, 2 Aug 2006, Jo Rhett wrote:
>> In the case of mismatched file daemons it's easy enough to work out which
>> one is playing up by running "status client" on each one.
>
> Nope. No configuration changes since I got that message, and no repeats of
> the message. "status client" worked on e
On Wed, 2 Aug 2006, Dan Langille wrote:
> Jo, quite simply, you are a damn pain in the ass. If a developer
> does not want to do the fix, there's nothing you can do. Leave it
> alone.
I've had this in as a _polite_ request to Kern for a couple of weeks, and
there's a feature request filed for
On Wed, 2 Aug 2006, Erik P. Olsen wrote:
>> Schedule {
>> Name = "EveryOtherTuesday"
>> Run = Level=Full tue w00,w02,w04,w06,w08,w10,w12(etc to
>> week 52) at 22:00
>> }
>
> Some years have 53 weeks!
Every year has part of a 53rd week.
A Calendar year is 52 weeks and one day
Nothing major
When backing up multiple sets on multiple clients, the current completion
messages aren't clear enough - especially on our SAN/NAS fileservers which
have upwards of 20Tb behind them and ~40 backups sets each.
The fix is simple.
In bacula-dir
Find:
mailcommand = "/usr/sb
On Thu, 3 Aug 2006, Kern Sibbald wrote:
>> I have a situation where we are about to add LTO3 drives to our changer,
>> which currently has LTO2 drives and media onboard.
>>
>> For obvious reasons I want to be able to use the LTO2 tapes in the LTO3
>> drives, but I also have to prevent any attempts
On Thu, 3 Aug 2006, Kern Sibbald wrote:
>> What's needed is the ability to define multiple mediatypes per tape drive.
>
> Unfortunately it is a lot more complicated than that, but 1.39 gets half way
> there.
That's good to hear.
I'm sure it get very complicated very quickly, but having achieved
On Fri, 4 Aug 2006, Jaime Ventura wrote:
> Ive installed the hardware and tested it using mtx command (from mtx
> OS package).
If you have this then you don't need the bacula-mtx package.
> When i try to uninstall mtx package, I get an error saying that
> mtx package is required by bacul
On Sat, 5 Aug 2006, Eric Bollengier wrote:
> You can try this (change Pool.Name and MediaType)
>
> SELECT Media.VolumeName AS volumename,
> Media.VolStatus AS volstatus,
> Media.LastWritten AS lastwritten,
> Media.MediaType AS mediatype,
> Media.VolMounts AS volmount
On Tue, 8 Aug 2006, Michel Meyers wrote:
> Those are mounted filesystems. Bacula will not descend into mounted
> filesystems unless you specifically tell it to.
Unfortunately it also doesn't backup the mountpoint as a directory, which
makes restoration a bit messy at times.
AB
---
On Thu, 10 Aug 2006, Jaime Ventura wrote:
> *mount
> The defined Storage resources are:
>1: File
>2: IBM Ultrium 3
Judging from this, you have malconfigured your autochanger resource.
Please re-read the manual and ensure you have an autochanger section AND
a section for each tape drive i
: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>
url: www.isep.ipp.pt <http://www.isep.ipp.pt>
Alan Brown wrote:
On Thu, 10 Aug 2006, Jaime Ventura wrote:
*mount
The defined Storage resources are:
1: File
2: IBM Ultrium 3
Judging from this, you have malconfigured yo
On Thu, 10 Aug 2006, Jeffrey L. Taylor wrote:
> On an unmodified Bacula system, dbcheck can run for days.
Literally.
> And it
> locks the database so no backups or restores can be done until it
> finishes. Look in the archives for this list about adding indexes.
> There are two or three indexe
On Thu, 10 Aug 2006 [EMAIL PROTECTED] wrote:
> => > On an unmodified Bacula system, dbcheck can run for days.
> => Literally.
>
> Ah. Ok. So, my attempts to run dbcheck were actually working normally.
Yes.
> => The queries for orphan files, paths and filenames will run ok once this is
> => done
On Fri, 11 Aug 2006, Dominic Marks wrote:
> I understand that dbcheck on a big database with millons and millions
> or rows like mine is going to take a long time. However, having the
> database locked for days on end is not really workable.
See my posting from yesterday. I have approx 100 millio
On Fri, 11 Aug 2006, Martin Simmons wrote:
> I'm thinking that it might be faster to restore a good copy of the database
> from a backup
Yes, IF you have a logfile you can replay too.
>. Don't forget that running dbcheck on a corrupted database
> might delete things that you still need (dependi
On Fri, 11 Aug 2006, Jaime Ventura wrote:
> Here they are.
You haven't defined the autochanger in your -dir file - you are addressing
the tape drive directly.
I'd advise you generate new passwords too - you just posted your existing
ones to the list.
>
> # Definition of file storage device
>
On Fri, 11 Aug 2006, Martin Simmons wrote:
>> Bacula. MySQL seems perfectly happy. Bacula has a number of symptoms which
>> make it unusable for a production system.
>
> Any idea what happened to cause the "More than one filename" problem? It
> would be nice to fix Bacula if possible.
More than
On Sun, 13 Aug 2006, Kern Sibbald wrote:
>> Unfortunately it also doesn't backup the mountpoint as a directory, which
>> makes restoration a bit messy at times.
>
> Bacula has backed up mount points for quite some time now, though I don't
> remember what version it was first implemented in.
It's
On Mon, 14 Aug 2006, Kern Sibbald wrote:
> On Friday 14 July 2006 15:36, Alan Brown wrote:
>>
>> Nothing major here.
>>
>> Kern:
>>
>> When doing a 'status storage' it would be extremely handy to know
>> whether jobs are:
>>
>&g
On Mon, 14 Aug 2006, Kern Sibbald wrote:
Unfortunately it also doesn't backup the mountpoint as a directory, which
makes restoration a bit messy at times.
>>>
>>> Bacula has backed up mount points for quite some time now, though I don't
>>> remember what version it was first implemented
On Mon, 14 Aug 2006, Frank Sweetser wrote:
> You can use the perror command to translate mysql error codes:
>
> tty/1 1011 erwin 11:45:48 $ perror 28
> OS error code 28: No space left on device
Or an oversize database.
The default maximum is about 4Gb for MyISAM.
You can set flags to make it
On Wed, 16 Aug 2006, Dirk Kleinhesselink wrote:
> On Wed, 16 Aug 2006, Kern Sibbald wrote:
>
> For what it's worth, I found that if I unloaded and re-loaded the
> SCSI controller, tape (st) and generic (sg) drivers on my system (linux,
> mandriva 2006) that would fix occasional problems with my ta
On Fri, 18 Aug 2006, Michael Brennen wrote:
> On Fri, 18 Aug 2006, Daniel Vesely wrote:
>
>> i have an VXA2 10U autochanger with 10 slots 1-10 but bacula think
>> that slots is 0-9 - so where i have setup that the slots start from
>> slot 1?
>
> My autochanger numbers from one, and Bacula is worki
On Wed, 23 Aug 2006, Magnus Hagander wrote:
>> I got it to work. I had to add the following line to
>> /etc/postgresql/8.1/main/pg_hba.conf:
>>
>> hostbacula bacula,root 127.0.0.1/32 trust
>>
>> Bacula is connecting with TCP/IP, not a Unix domain socket. This
>> was confusing me.
On Fri, 1 Sep 2006, Marco wrote:
> Just an idea:
>
> I did not build v1.38.11 and v1.38.8 myself but installed the debian
> packages. During the downgrade I noticed that there were changes concerning
> sqlite and sqlite3.
Sqlite is NOT a production quality database and should NOT be used in a
pr
On Fri, 1 Sep 2006, Arno Lehmann wrote:
>> 01-Sep 03:05 server-sd: Recycled volume "donderdag" on device "tape"
>> (/dev/st0), all previous data lost.
>> 01-Sep 05:19 server-sd: End of Volume "donderdag" at 3:6763 on device
>> "tape" (/dev/st0). Write of 64512 bytes got -1.
>> 01-Sep 05:19 server-
On Fri, 1 Sep 2006, Arno Lehmann wrote:
>> I'm backing up several servers but one specifically is about 86Gb data
>> and 1.6Million files. The data backup finish after a few hours (<6) but
>> then it (I assume) updates the database with all the attributes and that
>> takes >18h!
>> I looked a lit
(Kern, I'll file this in bugs shortly)
I've just found a couple of "gotchas" for those of us running multidrive
changers.
1: Running "update slots" against the changer device will only ever unload
drive 0
===
*update slots
The defined Storage resources are:
1: File
On Mon, 4 Sep 2006, Ferdinando Pasqualetti wrote:
> Too bad the disk-to-tape operation of spooling is done suspending the
> backup-to-disk operation (no concurrency with double buffering) and this
> limits the usefulness of spooling (as far as I know).
When multiple backups occur, they tend to
On Mon, 4 Sep 2006, Kern Sibbald wrote:
>> (Kern, I'll file this in bugs shortly)
>
> Probably not worth the effort. See my responses below.
> Bacula 1.38.11 and below do not permit doing a update slots at the same
> time a drive is in use.
Which is one reason update slots may have been requi
On Tue, 5 Sep 2006, Arno Lehmann wrote:
>> I get several messages like these on dmesg:
>> (scsi2:A:15:0): data overrun detected in Data-in phase. Tag == 0x3.
>> (scsi2:A:15:0): Have seen Data Phase. Length = 50. NumSGs = 1.
>> sg[0] - Addr 0x11c3b8000 : Length 50
>
> These are SCSI errors and i
On Wed, 6 Sep 2006, Jaime Ventura wrote:
> Although I made a kernel update, i checked the the data cable. Is was
> slightly disconnected. Probably that was the problem.
very likely.
> Now I have a tape with volStatus =Error.
> What now? What you recommend? delete the jobs associated with that ta
If memory serves me correctly, Bacula will not prune a volume which
contains the last remaining full backup of any given job.
Unfortunately I can't find the documentation for this feature. Can someone
remind me where it's lurking?
Thanks
Alan
-
On Fri, 29 May 2009, Olaf Zevenboom wrote:
> I can tell you that USB is a very very very bad idea to use / trust for
> reliable backups. Why? Pumping huge amounts of data over USB to a disk
> cause things to get very hot. Often to hot to handle resulting in
> BBQ-ing the USB-SATA converter or your
On Tue, 2 Jun 2009, Victor Sterpu wrote:
> I did restarted the bacula services, but still, the reported version is 2.4.4.
> I also installed the latest tar.gz 3.0.1 on another clean machine and the
> reported version is still 2.4.4.
try "which bacula-dir", "which bacula-sd", "which bacula-fd"
It
On Mon, 8 Jun 2009, Michael Hall wrote:
> Brand new Bacula user here. I'm having trouble finding up to date Bacula
> RPMs/SRPMs for CentOS 5, and have decided I may as well build my own.
CentOs5 = RHEL5 (and Scientific Linux, and Whitebox and a few other clones)
Use the RHEL5 rpm and all will be
On Wed, 22 Jul 2009, Bob Hetzel wrote:
>
> Has anybody tinkered around with spooling backups on an SSD (aka solid
> state drive) or a raid-0 pair of them for higher performance?
No, but it's on my TODO list (Trying to wrangle budget for them)
> It would seem that the issue of latency introduced
James Harper wrote:
> I'd hazard a guess that for spooling, raw throughput is more important
> than random access seek time, unless you spool is fragmented.
You arte correct - IF you're only spooling one backup.
I may have anything up to 6 running simultaneously, some people on this
list have m
On Fri, 24 Jul 2009, Marc Cousin wrote:
> All you really need is to be able to read and write big streams at the same
> time. So the real problem is to help your disk scheduler to be able to read
> while having a lot of data in the write cache.
How many streams are you handling simultaneously.
On Fri, 24 Jul 2009, John Lockard wrote:
> For spool, I would worry about the limited write (erase)
> cycles of SSD.
SLC drives are good for at least 100,000 cycles and all current SSD drives
use wear levelling plus even more generous levels of sparing than
mechanical HDDs. Even for constant spoo
601 - 700 of 1196 matches
Mail list logo
