Re: [Bacula-users] Bacula, Autochangers, insist on loading 'not in changer' media...
I would add that this problem appears to occur when all in-changer tapes
are unavailable due to volume status AND the job's pool does contain
available tapes, but those tapes are not in-changer. Bacula will attempt
to load a tape that is not in-changer, where it should send an operator
intervention notice.
On 5/12/23 06:22, Marco Gaiarin wrote:
We have some setups using bacula (debian buster, 9.4.2-2+deb10u1) and RDX
media, by the way of the 'vchanger' (Virtual Changer) script.
All works as expected, until the current mounted media exaust the 'in
changer' media (because exausted it; or because simply users load the
incorrect media...).
After that, bacula try to mount expired (purge them) or generically
avaliable volumes from other media, that are 'not in changer', putting them
on error.
We have extensivaly debugged vchanger script, that seems behave correctly.
Bacula seems have a current and correct state of 'in changer' volumes, and
anyway a 'update volumes' on console does not solve the trouble.
On director we have:
Autochanger {
Name = SDPVE2RDX
Address = sdpve2.sd.lnf.it
SDPort = 9103
Password = "unknown"
Maximum Concurrent Jobs = 5
Device = RDXAutochanger
Media Type = RDX
}
Pool {
Name = VEN-SD-SDPVE2RDXPool
Pool Type = Backup
Volume Use Duration = 1 days
Maximum Volume Jobs = 1
Recycle = yes
AutoPrune = yes
Action On Purge = Truncate
Volume Retention = 20 days
}
On SD we have:
Autochanger {
Name = RDXAutochanger
Device = RDXStorage0
Device = RDXStorage1
Device = RDXStorage2
Changer Command = "/usr/bin/vchanger %c %o %S %a %d"
Changer Device = "/etc/vchanger/SDPVE2RDX.conf"
}
Device {
Name = RDXStorage0
Drive Index = 0
Device Type = File
Media Type = RDX
RemovableMedia = no
RandomAccess = yes
Maximum Concurrent Jobs = 1
Archive Device = "/var/spool/vchanger/SDPVE2RDX/0"
}
Device {
Name = RDXStorage1
Drive Index = 1
Device Type = File
Media Type = RDX
RemovableMedia = no
RandomAccess = yes
Maximum Concurrent Jobs = 1
Archive Device = "/var/spool/vchanger/SDPVE2RDX/1"
}
Device {
Name = RDXStorage2
Drive Index = 2
Device Type = File
Media Type = RDX
RemovableMedia = no
RandomAccess = yes
Maximum Concurrent Jobs = 1
Archive Device = "/var/spool/vchanger/SDPVE2RDX/2"
}
Someone have some clue? Thanks.
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] compression with lto?
Gary R. Schmidt schrieb am 12.05.23 um 16:26: On 11/05/2023 13:49, Dr. Thorsten Brandau wrote: Hi I use an lto-9 drive for backup. I activated hardware compression on the tape (i think) but all the logs of bacula show a tape change at 18TB which should be the native (uncompressed) capacity. From experience I would expect a ratio of at last 1:1.2. Is there any way to make sure to use compression (or as last resort activate software compression)? Are you using the actual compressed device in your bacula-sd.conf? On Solaris I use /dev/rmt/0cbn to turn on compression. (But I started out handling tape back when they were huge upright things, throwing pseudo-random characters into device names is reflex. :-) ) Cheers, Gary B-) Hi Gary, well I remember times were you could change the whole disc with a cake basket like holder and tapes could be used as paperweights and through from the 5th floor, but will work fine afterwards... I am running linux and the tape is /dev/nst0 which should referr to the compressed device. The loader is sg7 but that should not play any role. I also remember times where tapes had so many names in /dev but also our older LTOs do not fuss around and leave little options on the dev side, but you have to configure them with MT or MTX. However, I turned compression on with MT and MTX, so I still do not understand why it seems that the data is imcompressible... I will try software compression to see IF the data is compressible (s many text files, should shrink nice an easy...). Cheers TB ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] compression with lto?
On 11/05/2023 13:49, Dr. Thorsten Brandau wrote: Hi I use an lto-9 drive for backup. I activated hardware compression on the tape (i think) but all the logs of bacula show a tape change at 18TB which should be the native (uncompressed) capacity. From experience I would expect a ratio of at last 1:1.2. Is there any way to make sure to use compression (or as last resort activate software compression)? Are you using the actual compressed device in your bacula-sd.conf? On Solaris I use /dev/rmt/0cbn to turn on compression. (But I started out handling tape back when they were huge upright things, throwing pseudo-random characters into device names is reflex. :-) ) Cheers, GaryB-) ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] compression with lto?
Bill Arlofski via Bacula-users schrieb am 11.05.23 um 20:45: On 5/10/23 21:49, Dr. Thorsten Brandau wrote: Hi I use an lto-9 drive for backup. I activated hardware compression on the tape (i think) but all the logs of bacula show a tape change at 18TB which should be the native (uncompressed) capacity. From experience I would expect a ratio of at last 1:1.2. Is there any way to make sure to use compression (or as last resort activate software compression)? Hello Thorsten, Of course, with LTO tape drives, you always want to enable hardware compression and disable software compression in Bacula. How that is set on your specific drive(s) I do not know. :) Bacula will just write until it gets an EOT (end of tape) from the tape drive, so the compression you will see is dependent on your dataset. Some tips: In your Director's configuration for the Storage resource which points to the SD's Tape drive (or library with multiple drives), you should set `AllowCompression = no` This will tell the FDs to not try to compress data before sending it to the SD when using this Storage resource in a job, even if you have a `compression = (lzo|gzip|zstd)` in your fileset. For jobs using this Storage, if compression in enabled for the fileset in use, you will see a harmless info log entry in the job logs telling you that "compression is not allowed for this storage device". (or similar) I prefer to enable/disable this in the Director's Storage resource and always enable compression in my filesets so that the same job and fileset, when using a different Director Storage that allows compression, it will automatically "Just Work"™ P.S. You can enable/disable compression in the SD's Devices, but I prefer to set it at what I would call the beginning of the chain in the Director's Storage resource. Home some of this is helpful, Bill Hi Bill, thank you. It still does not solve my problem why it does not compress the data on the tape... I think this is strange. I try to use GZIP6 now and see if the written data/total data changes then to see if I can safe some meters of tapes. And Backuptime. It seems that the Raid here is not the fastest to read and that makes the backup slow. Cheers T ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Bacula, Autochangers, insist on loading 'not in changer' media...
We have some setups using bacula (debian buster, 9.4.2-2+deb10u1) and RDX
media, by the way of the 'vchanger' (Virtual Changer) script.
All works as expected, until the current mounted media exaust the 'in
changer' media (because exausted it; or because simply users load the
incorrect media...).
After that, bacula try to mount expired (purge them) or generically
avaliable volumes from other media, that are 'not in changer', putting them
on error.
We have extensivaly debugged vchanger script, that seems behave correctly.
Bacula seems have a current and correct state of 'in changer' volumes, and
anyway a 'update volumes' on console does not solve the trouble.
On director we have:
Autochanger {
Name = SDPVE2RDX
Address = sdpve2.sd.lnf.it
SDPort = 9103
Password = "unknown"
Maximum Concurrent Jobs = 5
Device = RDXAutochanger
Media Type = RDX
}
Pool {
Name = VEN-SD-SDPVE2RDXPool
Pool Type = Backup
Volume Use Duration = 1 days
Maximum Volume Jobs = 1
Recycle = yes
AutoPrune = yes
Action On Purge = Truncate
Volume Retention = 20 days
}
On SD we have:
Autochanger {
Name = RDXAutochanger
Device = RDXStorage0
Device = RDXStorage1
Device = RDXStorage2
Changer Command = "/usr/bin/vchanger %c %o %S %a %d"
Changer Device = "/etc/vchanger/SDPVE2RDX.conf"
}
Device {
Name = RDXStorage0
Drive Index = 0
Device Type = File
Media Type = RDX
RemovableMedia = no
RandomAccess = yes
Maximum Concurrent Jobs = 1
Archive Device = "/var/spool/vchanger/SDPVE2RDX/0"
}
Device {
Name = RDXStorage1
Drive Index = 1
Device Type = File
Media Type = RDX
RemovableMedia = no
RandomAccess = yes
Maximum Concurrent Jobs = 1
Archive Device = "/var/spool/vchanger/SDPVE2RDX/1"
}
Device {
Name = RDXStorage2
Drive Index = 2
Device Type = File
Media Type = RDX
RemovableMedia = no
RandomAccess = yes
Maximum Concurrent Jobs = 1
Archive Device = "/var/spool/vchanger/SDPVE2RDX/2"
}
Someone have some clue? Thanks.
--
Ho ancora la forza di starvi a raccontare
le mie storie di sempre, di come posso amare (F. Guccini)
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
