Re: [Bacula-users] Autochanger does not change disk when full...
Hello, śr., 15 sty 2020 o 17:36 dmaziuk via Bacula-users < bacula-users@lists.sourceforge.net> napisał(a): > Use ZFS if you can: you get raid, lvm, and a filesystem in one coherent > package. > +1 :) -- Radosław Korzeniewski rados...@korzeniewski.net ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Autochanger does not change disk when full...
On 1/14/2020 11:31 PM, Stefan Melber wrote: Hi Josh, thanks for your explanation. Ok - looking on the pros and cons of all possibilities i will use finally an LVM over both RAID-arrays. I had already thought about that, but because i have to format both RAID-arrays for that (and therefore move some other stuff on that disks around) i first tried the autochanger... Use ZFS if you can: you get raid, lvm, and a filesystem in one coherent package. Dima ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Autochanger does not change disk when full...
Hi Josh, thanks for your explanation. Ok - looking on the pros and cons of all possibilities i will use finally an LVM over both RAID-arrays. I had already thought about that, but because i have to format both RAID-arrays for that (and therefore move some other stuff on that disks around) i first tried the autochanger... Best regards, Stefan On 1/14/2020 12:30 AM, Stefan Melber wrote: Hi Bacula-Users, i use a setup with two disks based on autochanger to save my backups - see configuration below. The changing between the two disks for different jobs works fine, but in the case during a backup one disk gets full, the backup fails because it does not change the disk in this case. So in my case, backup61which points on FileStorageRAID-Dev0 fills the disk up. Then bacula mounts a new volume (backup62), but instead of using FileStorageRAID-Dev1 is still uses (the full) FileStorageRAID-Dev0 which then fails because of disk full: Yes. Bacula assigns a Device, in this case FileStorageRAID-Dev0, to the job at startup. The Device assigned to the job can never change, once assigned. I'm not sure why this is the case, but it is by design, perhaps to prevent race conditions when asynchronous jobs are vying for drives and volumes. I'm not sure what happens when the same volume shows up in two different places, but it might be possible to create a symlink in RAID/BACKUP_RAID pointing to a volume file in RAID3/BACKUP_RAID. You could put both disk partitions in a LVM volume group and a single LV. The LV's single mountpoint would then be used as the ArchiveDevice for all of the Autochanger's Devices. Alternatively, you can use vchanger which treats all mountpoints as one large "tape" library and so allows any volume file in any mountpoint to be loaded into any of its drives (ie. associated Device resources). There are some trade-offs, such as vchanger doesn't support auto-labeling, so volumes must be manually added like when using a tape autoloader, but it does support barcode simulation, so creating and labeling volumes is not too onerous. 13-Jan 20:19 NASBox-dir JobId 153: Volume used once. Marking Volume "backup0061" as Used. 13-Jan 20:25 NASBox-dir JobId 153: Volume used once. Marking Volume "backup0061" as Used. 13-Jan 20:25 NASBox-sd JobId 153:*End of Volume "backup0061" at 3222547279824 on device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID). Write of 64512 bytes got -1.* 13-Jan 20:25 NASBox-sd JobId 153: End of medium on Volume "backup0061" Bytes=3,222,547,279,824 Blocks=49,952,680 at 13-Jan-2020 20:25. 13-Jan 20:25 NASBox-dir JobId 153: Created new Volume="backup0062", Pool="FileRAID", MediaType="FileRAID" in catalog. 13-Jan 20:25 NASBox-sd JobId 153: Labeled new Volume "backup0062" on File device "*FileStorageRAID-Dev0*" (/RAID/BACKUP_RAID). 13-Jan 20:25 NASBox-sd JobId 153: Wrote label to prelabeled Volume "backup0062" on File device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID) 13-Jan 20:25 NASBox-dir JobId 153: Volume used once. Marking Volume "backup0062" as Used. 13-Jan 20:25 NASBox-sd JobId 153: New volume "backup0062" mounted on device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID) at 13-Jan-2020 20:25. 13-Jan 20:27 NASBox-sd JobId 153: Fatal error: Out of freespace caused End of Volume "backup0062" at 774595809 on device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID). Write of 64512 bytes got -1. 13-Jan 20:27 bsastf670-fd JobId 153: Error: bsock.c:646 Write error sending 65540 bytes to Storage daemon:NASBox:9103: ERR=Connection reset by peer 13-Jan 20:27 NASBox-sd JobId 153: Elapsed time=13:41:01, Transfer rate=65.38 M Bytes/second 13-Jan 20:27 bsastf670-fd JobId 153: Fatal error: backup.c:842 Network send error to SD. ERR=Connection reset by peer 13-Jan 20:27 bsastf670-fd JobId 153: Error: bsock.c:536 Socket has errors=1 on call to Storage daemon:NASBox:9103 13-Jan 20:28 NASBox-dir JobId 153: Error: Bacula NASBox-dir 9.0.6 (20Nov17): How can i prevent this and give bacula a hint to change to the free disk? Best regards, Stefan Autochanger { Name = FileStorageRAID Device = FileStorageRAID-Dev0, FileStorageRAID-Dev1 Changer Command = /dev/null Changer Device = /dev/null } Device { Name = FileStorageRAID-Dev0 Media Type = FileRAID Archive Device = /RAID/BACKUP_RAID LabelMedia = yes Random Access = yes AutomaticMount = yes RemovableMedia = no AlwaysOpen = no Maximum Concurrent Jobs = 5 Autochanger = yes Drive Index = 0 } Device { Name = FileStorageRAID-Dev1 Media Type = FileRAID Archive Device = /RAID3/BACKUP_RAID LabelMedia = yes Random Access = yes AutomaticMount = yes RemovableMedia = no AlwaysOpen = no Maximum Concurrent Jobs = 5 Autochanger = yes Drive Index = 1 } ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net
Re: [Bacula-users] Autochanger does not change disk when full...
On 1/14/2020 12:30 AM, Stefan Melber wrote: Hi Bacula-Users, i use a setup with two disks based on autochanger to save my backups - see configuration below. The changing between the two disks for different jobs works fine, but in the case during a backup one disk gets full, the backup fails because it does not change the disk in this case. So in my case, backup61which points on FileStorageRAID-Dev0 fills the disk up. Then bacula mounts a new volume (backup62), but instead of using FileStorageRAID-Dev1 is still uses (the full) FileStorageRAID-Dev0 which then fails because of disk full: Yes. Bacula assigns a Device, in this case FileStorageRAID-Dev0, to the job at startup. The Device assigned to the job can never change, once assigned. I'm not sure why this is the case, but it is by design, perhaps to prevent race conditions when asynchronous jobs are vying for drives and volumes. I'm not sure what happens when the same volume shows up in two different places, but it might be possible to create a symlink in RAID/BACKUP_RAID pointing to a volume file in RAID3/BACKUP_RAID. You could put both disk partitions in a LVM volume group and a single LV. The LV's single mountpoint would then be used as the ArchiveDevice for all of the Autochanger's Devices. Alternatively, you can use vchanger which treats all mountpoints as one large "tape" library and so allows any volume file in any mountpoint to be loaded into any of its drives (ie. associated Device resources). There are some trade-offs, such as vchanger doesn't support auto-labeling, so volumes must be manually added like when using a tape autoloader, but it does support barcode simulation, so creating and labeling volumes is not too onerous. 13-Jan 20:19 NASBox-dir JobId 153: Volume used once. Marking Volume "backup0061" as Used. 13-Jan 20:25 NASBox-dir JobId 153: Volume used once. Marking Volume "backup0061" as Used. 13-Jan 20:25 NASBox-sd JobId 153:*End of Volume "backup0061" at 3222547279824 on device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID). Write of 64512 bytes got -1.* 13-Jan 20:25 NASBox-sd JobId 153: End of medium on Volume "backup0061" Bytes=3,222,547,279,824 Blocks=49,952,680 at 13-Jan-2020 20:25. 13-Jan 20:25 NASBox-dir JobId 153: Created new Volume="backup0062", Pool="FileRAID", MediaType="FileRAID" in catalog. 13-Jan 20:25 NASBox-sd JobId 153: Labeled new Volume "backup0062" on File device "*FileStorageRAID-Dev0*" (/RAID/BACKUP_RAID). 13-Jan 20:25 NASBox-sd JobId 153: Wrote label to prelabeled Volume "backup0062" on File device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID) 13-Jan 20:25 NASBox-dir JobId 153: Volume used once. Marking Volume "backup0062" as Used. 13-Jan 20:25 NASBox-sd JobId 153: New volume "backup0062" mounted on device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID) at 13-Jan-2020 20:25. 13-Jan 20:27 NASBox-sd JobId 153: Fatal error: Out of freespace caused End of Volume "backup0062" at 774595809 on device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID). Write of 64512 bytes got -1. 13-Jan 20:27 bsastf670-fd JobId 153: Error: bsock.c:646 Write error sending 65540 bytes to Storage daemon:NASBox:9103: ERR=Connection reset by peer 13-Jan 20:27 NASBox-sd JobId 153: Elapsed time=13:41:01, Transfer rate=65.38 M Bytes/second 13-Jan 20:27 bsastf670-fd JobId 153: Fatal error: backup.c:842 Network send error to SD. ERR=Connection reset by peer 13-Jan 20:27 bsastf670-fd JobId 153: Error: bsock.c:536 Socket has errors=1 on call to Storage daemon:NASBox:9103 13-Jan 20:28 NASBox-dir JobId 153: Error: Bacula NASBox-dir 9.0.6 (20Nov17): How can i prevent this and give bacula a hint to change to the free disk? Best regards, Stefan Autochanger { Name = FileStorageRAID Device = FileStorageRAID-Dev0, FileStorageRAID-Dev1 Changer Command = /dev/null Changer Device = /dev/null } Device { Name = FileStorageRAID-Dev0 Media Type = FileRAID Archive Device = /RAID/BACKUP_RAID LabelMedia = yes Random Access = yes AutomaticMount = yes RemovableMedia = no AlwaysOpen = no Maximum Concurrent Jobs = 5 Autochanger = yes Drive Index = 0 } Device { Name = FileStorageRAID-Dev1 Media Type = FileRAID Archive Device = /RAID3/BACKUP_RAID LabelMedia = yes Random Access = yes AutomaticMount = yes RemovableMedia = no AlwaysOpen = no Maximum Concurrent Jobs = 5 Autochanger = yes Drive Index = 1 } ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Autochanger does not change disk when full...
Hi Bacula-Users, i use a setup with two disks based on autochanger to save my backups - see configuration below. The changing between the two disks for different jobs works fine, but in the case during a backup one disk gets full, the backup fails because it does not change the disk in this case. So in my case, backup61which points on FileStorageRAID-Dev0 fills the disk up. Then bacula mounts a new volume (backup62), but instead of using FileStorageRAID-Dev1 is still uses (the full) FileStorageRAID-Dev0 which then fails because of disk full: 13-Jan 20:19 NASBox-dir JobId 153: Volume used once. Marking Volume "backup0061" as Used. 13-Jan 20:25 NASBox-dir JobId 153: Volume used once. Marking Volume "backup0061" as Used. 13-Jan 20:25 NASBox-sd JobId 153:*End of Volume "backup0061" at 3222547279824 on device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID). Write of 64512 bytes got -1.* 13-Jan 20:25 NASBox-sd JobId 153: End of medium on Volume "backup0061" Bytes=3,222,547,279,824 Blocks=49,952,680 at 13-Jan-2020 20:25. 13-Jan 20:25 NASBox-dir JobId 153: Created new Volume="backup0062", Pool="FileRAID", MediaType="FileRAID" in catalog. 13-Jan 20:25 NASBox-sd JobId 153: Labeled new Volume "backup0062" on File device "*FileStorageRAID-Dev0*" (/RAID/BACKUP_RAID). 13-Jan 20:25 NASBox-sd JobId 153: Wrote label to prelabeled Volume "backup0062" on File device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID) 13-Jan 20:25 NASBox-dir JobId 153: Volume used once. Marking Volume "backup0062" as Used. 13-Jan 20:25 NASBox-sd JobId 153: New volume "backup0062" mounted on device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID) at 13-Jan-2020 20:25. 13-Jan 20:27 NASBox-sd JobId 153: Fatal error: Out of freespace caused End of Volume "backup0062" at 774595809 on device "FileStorageRAID-Dev0" (/RAID/BACKUP_RAID). Write of 64512 bytes got -1. 13-Jan 20:27 bsastf670-fd JobId 153: Error: bsock.c:646 Write error sending 65540 bytes to Storage daemon:NASBox:9103: ERR=Connection reset by peer 13-Jan 20:27 NASBox-sd JobId 153: Elapsed time=13:41:01, Transfer rate=65.38 M Bytes/second 13-Jan 20:27 bsastf670-fd JobId 153: Fatal error: backup.c:842 Network send error to SD. ERR=Connection reset by peer 13-Jan 20:27 bsastf670-fd JobId 153: Error: bsock.c:536 Socket has errors=1 on call to Storage daemon:NASBox:9103 13-Jan 20:28 NASBox-dir JobId 153: Error: Bacula NASBox-dir 9.0.6 (20Nov17): How can i prevent this and give bacula a hint to change to the free disk? Best regards, Stefan Autochanger { Name = FileStorageRAID Device = FileStorageRAID-Dev0, FileStorageRAID-Dev1 Changer Command = /dev/null Changer Device = /dev/null } Device { Name = FileStorageRAID-Dev0 Media Type = FileRAID Archive Device = /RAID/BACKUP_RAID LabelMedia = yes Random Access = yes AutomaticMount = yes RemovableMedia = no AlwaysOpen = no Maximum Concurrent Jobs = 5 Autochanger = yes Drive Index = 0 } Device { Name = FileStorageRAID-Dev1 Media Type = FileRAID Archive Device = /RAID3/BACKUP_RAID LabelMedia = yes Random Access = yes AutomaticMount = yes RemovableMedia = no AlwaysOpen = no Maximum Concurrent Jobs = 5 Autochanger = yes Drive Index = 1 } ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users