Re: [Bacula-users] Job pre-allocating 20 volumes from a pool un-necessary
2010-11-29 05:20:33
You can build 5.0.3 on Debian Squeeze from source, its so easy.
The wiki has 'from source' howtos on Debian that could help you.
Javier. 2010/11/29 Francois Duchatelet <bacula-f AT duchatelet DOT org>
No it isn’t.
The normal way of working is:
- fill a volume,
- THEN look for an available (Appendable,Purged,...) one,
- THEN ONLY if there are no available volumes in the Pool, create a new one
(if the pool allows it).
The issue is: SOMETIMES, the locutus job PRE-creates 20 volumes, although
it doesn’t need them.
In this case, the job was only 400Mb and normally completed on ONE
volume.
The “Sleep 60” is necessary because otherwise the 2 jobs, starting at the
same, would compete for the same volume.
With this ‘sleep 60’, the sulu job has time to mark it’s first volume as
‘Used’ (by virtue of ‘Maximum
Volume Jobs = 1‘) before the locutus job does its
reservation.
At least this was an issue before I upgraded to
5.0.2, while the two jobs ran concurrently.
As I need to upgrade to 5.0.3 anyway, I’ll see if
the problem persist.
Does someone have 5.0.3 packages for
debian/squeeze ? Otherwize I’ll try to build them.
Cheers,
François
Sent: Sunday, November 28, 2010 9:45 PM
Subject: Re: [Bacula-users] Job pre-allocating 20 volumes from a
poolun-necessary
It
seems correct! How many GB you have on spock host? You difined maximum
volume size with 1GB and only one use for each volume, then if the backup of
spock host exceed 1GB other volumes will be created. If you want keep the job on
single volume you should remove the maximum volume size definition. Why
you need this "sleep 60"? Kleber
2010/11/28 Francois D. <bacula-f AT duchatelet DOT org>
Hello,
I've an issue preventing the correct recycling
of my Pool's volume.
Given the definitions at the end of the message, I
have the job for locutus pre-create 20 volumes in the Default pool right
after the 60 seconds sleep.
What happens: - at 02:05, two jobs (Sulu
and Locutus) starts. - to avoid them compelling for the same Volume, the
Locutus job does a 'sleep 60'. - due, I think, to the known bug in 5.0.2
(bug-1578), the locutus job waits for the sulu completion before going
on. - sometimes (I was not able to deduce a pattern yet), while it is
waiting, the locutus job creates new volumes in the target pool, even
before it needs it.
When the pool is empty, like today (I've cleaned
and rebuild the pool yesterday), this is not really an issue, but when
there are 100+ volumes, with about 50 of them purged, and ready to be
re-used, this is annoying, as this makes the pool growing.
Any ideas
of what I did wrong
?
Cheers, François
=============================================
Here
is an extract of the Sulu job log: -------------------- 28-Nov 02:05
spock-dir JobId 1446: Start Backup JobId
1446, Job=sulu.2010-11-28_02.05.00_44 28-Nov 02:05 spock-dir JobId 1446:
Created new Volume "BDSK-002" in catalog. 28-Nov 02:05 spock-dir JobId
1446: Using Device "FileStorage" 28-Nov 02:05 spock-sd JobId 1446: Labeled
new Volume "BDSK-002" on device "FileStorage" (/backups). 28-Nov 02:05
spock-sd JobId 1446: Wrote label to prelabeled Volume "BDSK-002" on device
"FileStorage" (/backups) 28-Nov 02:05 spock-dir JobId 1446: Max Volume jobs
exceeded. Marking Volume "BDSK-002" as Used. 28-Nov 02:05 sulu-fd JobId
1446: Generate VSS snapshots. Driver="VSS Vista", Drive(s)="C" 28-Nov
02:08 spock-sd JobId 1446: User defined maximum volume
capacity 1,073,741,824 exceeded on device "FileStorage"
(/backups). 28-Nov 02:08 spock-sd JobId 1446: End of medium on Volume
"BDSK-002" Bytes=1,073,737,912 Blocks=16,644 at 28-Nov-2010
02:08. 28-Nov 02:08 spock-sd JobId 1446: Labeled new Volume "BDSK-003" on
device "FileStorage" (/backups). 28-Nov 02:08 spock-sd JobId 1446: Wrote
label to prelabeled Volume "BDSK-003" on device "FileStorage"
(/backups) 28-Nov 02:08 spock-dir JobId 1446: Max Volume jobs exceeded.
Marking Volume "BDSK-003" as Used. 28-Nov 02:08 spock-sd JobId 1446: New
volume "BDSK-003" mounted on device "FileStorage" (/backups) at 28-Nov-2010
02:08. 28-Nov 02:24 spock-sd JobId 1446: User defined maximum volume
capacity 1,073,741,824 exceeded on device "FileStorage"
(/backups). 28-Nov 02:24 spock-sd JobId 1446: End of medium on Volume
"BDSK-003" Bytes=1,073,737,675 Blocks=16,644 at 28-Nov-2010
02:24. 28-Nov 02:24 spock-sd JobId 1446: Labeled new Volume "BDSK-004" on
device "FileStorage" (/backups). 28-Nov 02:24 spock-sd JobId 1446: Wrote
label to prelabeled Volume "BDSK-004" on device "FileStorage"
(/backups) 28-Nov 02:24 spock-dir JobId 1446: Max Volume jobs exceeded.
Marking Volume "BDSK-004" as Used. 28-Nov 02:24 spock-sd JobId 1446: New
volume "BDSK-004" mounted on device "FileStorage" (/backups) at 28-Nov-2010
02:24.
And here is an extract of the locutus job
log: -------------------- 28-Nov 02:05 spock-dir JobId 1447: shell
command: run BeforeJob "/bin/sleep 60" 28-Nov 02:06 spock-dir JobId
1447: Start Backup JobId 1447, Job=locutus.2010-11-28_02.05.01_45 28-Nov
02:06 spock-dir JobId 1447: Created new Volume "BDSK-003" in
catalog. 28-Nov 02:06 spock-dir JobId 1447: Created new Volume "BDSK-004"
in catalog. 28-Nov 02:06 spock-dir JobId 1447: Created new Volume
"BDSK-005" in catalog. 28-Nov 02:06 spock-dir JobId 1447: Created new
Volume "BDSK-006" in catalog. 28-Nov 02:06 spock-dir JobId 1447: Created
new Volume "BDSK-007" in catalog. 28-Nov 02:06 spock-dir JobId 1447:
Created new Volume "BDSK-008" in catalog. 28-Nov 02:06 spock-dir JobId
1447: Created new Volume "BDSK-009" in catalog. 28-Nov 02:06 spock-dir
JobId 1447: Created new Volume "BDSK-010" in catalog. 28-Nov 02:06
spock-dir JobId 1447: Created new Volume "BDSK-011" in catalog. 28-Nov
02:06 spock-dir JobId 1447: Created new Volume "BDSK-012" in
catalog. 28-Nov 02:06 spock-dir JobId 1447: Created new Volume "BDSK-013"
in catalog. 28-Nov 02:06 spock-dir JobId 1447: Created new Volume
"BDSK-014" in catalog. 28-Nov 02:06 spock-dir JobId 1447: Created new
Volume "BDSK-015" in catalog. 28-Nov 02:06 spock-dir JobId 1447: Created
new Volume "BDSK-016" in catalog. 28-Nov 02:06 spock-dir JobId 1447:
Created new Volume "BDSK-017" in catalog. 28-Nov 02:06 spock-dir JobId
1447: Created new Volume "BDSK-018" in catalog. 28-Nov 02:06 spock-dir
JobId 1447: Created new Volume "BDSK-019" in catalog. 28-Nov 02:06
spock-dir JobId 1447: Created new Volume "BDSK-020" in catalog. 28-Nov
02:06 spock-dir JobId 1447: Created new Volume "BDSK-021" in
catalog. 28-Nov 02:08 spock-dir JobId 1447: Created new Volume "BDSK-022"
in catalog. 28-Nov 02:24 spock-dir JobId 1447: Created new Volume
"BDSK-023" in catalog. 28-Nov 02:29 spock-dir JobId 1447: Using Device
"FileStorage" 28-Nov 02:29 spock-sd JobId 1447: Labeled new Volume
"BDSK-005" on device "FileStorage" (/backups). 28-Nov 02:29 spock-sd
JobId 1447: Wrote label to prelabeled Volume "BDSK-005" on device
"FileStorage" (/backups)
Extract of
configuration: ------------------------- # Default pool definition,
targeting the disk storage Pool { Name = Default Pool Type =
Backup Recycle =
yes
# Bacula can automatically recycle Volumes AutoPrune =
yes
# Prune expired volumes Volume Retention = 7 days Maximum Volume Jobs =
1 # On disk, I want a new volume for each
job Maximum Volume Bytes = 1G # Limit the size of
volumes to something reasonnable for disk files LabelFormat =
"BDSK-${BDSK+:p/3/0/r}" Storage = File Next Pool = LTO1 Migration
Time = 1 hour }
# Those are generic defaults for all my
backups JobDefs { Name = "DefaultJob" Type = Backup Level =
Incremental Schedule = "WeeklyCycle" Messages = Standard Pool =
Default Priority = 10 }
Job { Name = "sulu" Client =
sulu-fd FileSet = "Sulu Set" JobDefs = "DefaultJob" # Force it to be
started ahead of others, so the schedule is kept" Priority =
9 }
Job { Name = "locutus" Client = locutus-fd FileSet =
"Locutus Set" JobDefs = "DefaultJob" # Force it to be started ahead of
others, so the schedule is kept" Priority = 9
RunScript
{ RunsWhen = Before RunsOnClient =
No FailJobOnError = No Command = "/bin/sleep
60" } }
------------------------------------------------------------------------------ Increase
Visibility of Your 3D Game App & Earn a Chance To Win $500! Tap into
the largest installed PC base & get more eyes on your game
by optimizing for Intel(R) Graphics Technology. Get started today with
the Intel(R) Software Partner Program. Five $500 cash prizes are up for
grabs. http://p.sf.net/sfu/intelisp-dev2dev _______________________________________________ Bacula-users
mailing list Bacula-users AT lists.sourceforge DOT net https://lists.sourceforge.net/lists/listinfo/bacula-users
------------------------------------------------------------------------------
Increase Visibility of Your 3D Game App & Earn a Chance To Win $500!
Tap into the largest installed PC base & get more eyes on your game by
optimizing for Intel(R) Graphics Technology. Get started today with the
Intel(R) Software Partner Program. Five $500 cash prizes are up for grabs.
http://p.sf.net/sfu/intelisp-dev2dev _______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
-- _____________________________________________ Francisco Javier Funes Nieto [ esencia AT gmail DOT com] CANONIGOS Servicios Informáticos para PYMES.
Cl. Cruz 2, 1º Oficina 7 Tlf: 958.536759 / 661134556 Fax: 958.521354 GRANADA - 18002
------------------------------------------------------------------------------
Increase Visibility of Your 3D Game App & Earn a Chance To Win $500!
Tap into the largest installed PC base & get more eyes on your game by
optimizing for Intel(R) Graphics Technology. Get started today with the
Intel(R) Software Partner Program. Five $500 cash prizes are up for grabs.
http://p.sf.net/sfu/intelisp-dev2dev _______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|
|
|