On Wednesday, August 31, 2016 at 3:02:52 PM UTC-5, Ruth Ivimey-Cook wrote:
> I have bareos backing up to two LTO tape drives, and occasionally see errors
> such as the following. This problem is not specific to a particular client. I
> don't think it's specific to a particular drive, though mostly it happens
> with the LTO3 rather than the LTO4. Howver The LTO3 gets more use...
>
> I'd really like a solution because these errors require manual intervention
> to fix, wasting my time and also hold up the backup process for everything.
>
> ... earlier messages
> 2016-08-31 19:40:56 helva-sd JobId 1088: Despooling elapsed time =
> 00:23:49, Transfer rate = 89.57 M Bytes/second
> 2016-08-31 19:40:56 helva-sd JobId 1088: Fatal error: spool.c:306 Fatal
> append error on device "LTO3TapeStorage" (/dev/tape/by-id/scsi-3500110a-nst):
> ERR=block.c:1003 Read error on fd=4 at file:blk 0:0 on device
> "LTO3TapeStorage" (/dev/tape/by-id/scsi-3500110a-nst). ERR=Cannot allocate
> memory.
> 2016-08-31 19:40:56 helva-sd JobId 1088: Fatal error: Re-read of last
> block: block numbers differ by more than one.
> Probable tape misconfiguration and data loss. Read block=0 Want block=178510.
>
>
> Director and Storage are on same machine. The job messages above were from a
> remote client but the data is spooled via SSD. Drive I/O typically achieves
> 90MB/s.
>
> The error messages suggest to me that the problem is not in fact with the
> tape drive but something else. Given the previous append error, I suspect the
> attempt to read the previous block is not in fact reading the previous block
> but the "current" failed one, somehow.
>
> If the message text is correct and it is a tape misconfiguration, what sorts
> of things might be wrong?
>
> Of course, I'm a complete novice at this. Can anyone help identify what is
> really happening?
>
> Might it be possible to modify the code to skip back several (e.g. 4) blocks
> and reread/verify them all, and then carry on writing (possibly on the next
> tape) from one after the last known 'good' block?
>
> Regards
> Ruth
>
>
> Config details:
>
> Intel i5-2500K at 3.3GHz with 8GB RAM, 16GB swap.
> One HP LTO3 tape drive connected using U320-SCSI on Symbios 53c1030 PCIe
> interface at 160MB/s. Firmware D26D <== drive involved in error above.
> One Quantum LTO4 tape drive connected using U320-SCSI on Symbios 53c1030 PCIe
> interface at 320MB/s. Firmware W53F <== connected to same scsi bus
>
> Ubuntu 16.04 LTS, kernel 4.4.0-34, very lightly loaded.
> Bareos installed from Bareos.org repo, version
> 16.3.1.1471011706.f4784a7-1304.1
> Volume data spooling from Samsung SSD on SATA-II interface.
>
> SD config for the LTO3 drive:
> Device {
> Name = LTO3TapeStorage
> Drive Index = 1
> Autochanger = no
> Archive Device = /dev/tape/by-id/scsi-3500110a-nst
> AutomaticMount = yes
> AlwaysOpen = yes
> Media Type = LTO3
> RandomAccess = no
> RemovableMedia = yes
> Offline On Unmount = yes
> Spool Directory = /media/tapespool/LTO3
> Maximum Block Size = 512kb
> Maximum File Size = 32gb
> Maximum Network Buffer Size = 65536
> Maximum Concurrent Jobs = 1
> }
your block size is wrong. The tape block size should be 512b not 512kb. 512 is
small though you might want to try 1024b
--
You received this message because you are subscribed to the Google Groups
"bareos-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
For more options, visit https://groups.google.com/d/optout.
