Le 25/06/2018 à 11:37, Jérôme Blion a écrit :
Hello everyone,
I have a Bacula 7.4.4 installed on a Debian Stretch. I set it up when
it was a 5.2.6 version. I want to have a 35 days retention for "flat
files". I set File retention and Job retention to 35 days on each
client declared.
Client {
Name = ...-fd
Address = ...
FDPort = 9102
Catalog = Catalog1
Password = "..." # password for FileDaemon
File Retention = 35 days # 1 Month
Job Retention = 35 days # 1 Month
AutoPrune = yes # Prune expired Jobs/Files
Maximum Concurrent Jobs = 20
}
To restore a 30 days old file, I need at least one full backup that
may be older than 30 days. To keep 2 full backup, I set up volume
Retention to 70 days:
Pool {
Name = POL-35D
Pool Type = Backup
Recycle = yes # Bacula can automatically
recycle Volumes
LabelFormat = "Vol"
AutoPrune = yes # Prune expired volumes
Volume Retention = 70 days # one month
Maximum Volume Bytes = 50G # Limit Volume size to something
reasonable
Maximum Volumes = 200 # Limit number of Volumes in Pool
Recycle Oldest Volume = yes
RecyclePool = Scratch
}
I used to have Volume Retention set to 35 days before, I changed that
recently. Today, I checked in the PostgreSQL database:
bacula=# select volumename, pool.name, volstatus, media.enabled,
media.recycle, media.actiononpurge, media.volretention from media
bacula-# left join pool using(poolid)
bacula-# where pool.name='POL-35D';
volumename | name | volstatus | enabled | recycle | actiononpurge
| volretention
------------+---------+-----------+---------+---------+---------------+--------------
Vol0083 | POL-35D | Full | 1 | 1 | 1
| 6048000
Vol0053 | POL-35D | Full | 1 | 1 | 1
| 6048000
[...]
Vol0036 | POL-35D | Append | 1 | 1 | 0
| 6048000
[...]
Vol0054 | POL-35D | Full | 1 | 1 | 1
| 6048000
(54 lignes)
Any volume in the pool has the same volretention : 70 days.
Here is the schedule I use:
Schedule {
Name = "WeeklyCycle"
Run = Full 1st sat at 22:05
Run = Differential 2nd-5th sat at 22:05
Run = Incremental sun-fri at 22:05
}
The problem is the following: The full backup is pruned too quickly.
To restore a 30 days old, sometimes, I need a full older than 30 days
(to use the differential then the incremental I need)
Here is what I have for a specific client:
bacula=# select job, type, level, starttime, jobtdate from job
bacula-# where name='Backup_carina'
bacula-# order by schedtime;
job | type | level | starttime
| jobtdate
--------------------------------------+------+-------+---------------------+------------
Backup_carina.2018-05-20_12.05.00_47 | B | I | 2018-05-20
12:26:00 | 1526812010
Backup_carina.2018-05-21_12.05.00_24 | B | I | 2018-05-21
12:26:30 | 1526898461
Backup_carina.2018-05-22_12.05.00_21 | B | I | 2018-05-22
12:35:49 | 1526985397
Backup_carina.2018-05-23_12.05.00_08 | B | I | 2018-05-23
12:31:07 | 1527071535
Backup_carina.2018-05-24_12.05.00_40 | B | I | 2018-05-24
12:05:02 | 1527156349
Backup_carina.2018-05-25_12.05.00_12 | B | I | 2018-05-25
12:05:02 | 1527242745
Backup_carina.2018-05-26_12.05.00_44 | B | D | 2018-05-26
12:05:03 | 1527329230
Backup_carina.2018-05-27_12.05.00_17 | B | I | 2018-05-27
12:05:03 | 1527415549
Backup_carina.2018-05-28_12.05.00_49 | B | I | 2018-05-28
12:05:03 | 1527501962
Backup_carina.2018-05-29_12.05.00_22 | B | I | 2018-05-29
12:32:24 | 1527589991
Backup_carina.2018-05-30_12.05.00_00 | B | I | 2018-05-30
12:05:02 | 1527674746
Backup_carina.2018-05-31_12.05.00_30 | B | I | 2018-05-31
12:05:03 | 1527761144
Backup_carina.2018-06-01_12.05.00_00 | B | I | 2018-06-01
12:05:02 | 1527847545
Backup_carina.2018-06-02_12.05.00_30 | B | F | 2018-06-02
12:18:09 | 1527935475
Backup_carina.2018-06-03_12.05.00_00 | B | I | 2018-06-03
12:05:02 | 1528020348
Backup_carina.2018-06-04_12.05.00_30 | B | I | 2018-06-04
12:05:02 | 1528106747
Backup_carina.2018-06-05_12.05.00_34 | B | I | 2018-06-05
12:15:52 | 1528193799
[...]
The full backup worked fine at beginning of June. But the May's full
backup has been pruned. I can not restore data pieces from backups
prior to 2 of June. I would need at least one full and one
differential to: have something working.
I am wrong somewhere but I can't identify what I did wrong.
Do you have some things I could investigate to identify why my full
backups are pruned too quickly ?
Kind regards.
Jérôme Blion.
Hello,
It's me again. I tried a different thing. In my client definition:
File Retention = 70 days
Job Retention = 70 days
==> This way, catalog is not cleaned too early anymore. According to the
documentation, I should use these 2 retention periods only to avoid the
catalog to grow too much. It should not be used to manage the backup
retention.
In my Job definition:
Pool = POL-35D
Full Backup Pool = POL-70D
Differential Backup Pool = POL-42D
Incremental Backup Pool = POL-35D
Each type of backup will go to a different pool with an appropriate
volume retention.
I don't think modifying schedules to override Full/Diff/Inc pools is a
good idea as I could want to run a manual backup and use the correct pool.
I'll check next month if everything goes fine.
HTH.
Jérôme Blion.
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
