When I try to Restore/Migrate a specific job distributed on two tapes, the
first tape is read successfully but when mounting the second tape the job
fails with error
```
20-Jul 08:02 bareos-sd JobId 21833: Ready to read from volume "NIX417L6" on
device "tapedrive-0" (/dev/tape/by-id/scsi-350050763121a063c-nst).
20-Jul 08:02 bareos-sd JobId 21833: Forward spacing Volume "NIX417L6" to
file:block 0:1.
20-Jul 08:02 bareos-sd JobId 21833: Error: stored/block.cc:1057 Read error
on fd=7 at file:blk 0:1 on device "tapedrive-0"
(/dev/tape/by-id/scsi-350050763121a063c-nst). ERR=Cannot allocate memory.
```
This only happens on this specific tape. Other jobs distributed on multiple
tapes work correctly.
I did configure the Bareos Tape device
/etc/bareos/bareos-sd.d/device/tapedrive-0.conf with "Maximum Block Size =
1048576" from the very beginning and all tapes were written using this
configuration.
Also all relevant pools have "Maximum Block Size = 1048576" set.
I started using Bareos after Version 14.2.0 so the new block size handling
(label block 63k, data blocks 1M) was used on all tapes.
But the error suggests that there is a problem with the block size.
This is seconded by /var/log/messages
```
Jul 20 08:02:14 igms07 kernel: st 3:0:0:0: [st0] Failed to read 1048576
byte block with 64512 byte transfer.
```
So it seems that Bareos is trying to read with 63k block size although it
is configured otherwise?
Further tests showed that I can read the tape label correctly with btape:
```
]$ btape /dev/tape/by-id/scsi-350050763121a063c-nst
Tape block granularity is 1024 bytes.
btape: stored/butil.cc:290-0 Using device:
"/dev/tape/by-id/scsi-350050763121a063c-nst" for writing.
btape: stored/btape.cc:485-0 open device "tapedrive-0"
(/dev/tape/by-id/scsi-350050763121a063c-nst): OK
*status
Bareos status: file=0 block=0
Device status: BOT ONLINE IM_REP_EN file=0 block=0
Device status: TAPE BOT ONLINE IMMREPORT. ERR=
*readlabel
btape: stored/btape.cc:532-0 Volume label read correctly.
Volume Label:
Id : Bareos 2.0 immortal
VerNo : 20
VolName : NIX417L6
PrevVolName :
VolFile : 0
LabelType : VOL_LABEL
LabelSize : 167
PoolName : Scratch
MediaType : LTO
PoolType : Backup
HostName : igms07.vision.local
Date label written: 05-Nov-2018 11:42
```
I can read the tape with dd with a blocksize of 1M:
dd if=/dev/tape/by-id/scsi-350050763121a063c-nst bs=1M
btape scan fails with the same error as Bareos:
```
]$ btape /dev/tape/by-id/scsi-350050763121a063c-nst
Tape block granularity is 1024 bytes.
btape: stored/butil.cc:290-0 Using device:
"/dev/tape/by-id/scsi-350050763121a063c-nst" for writing.
btape: stored/btape.cc:485-0 open device "tapedrive-0"
(/dev/tape/by-id/scsi-350050763121a063c-nst): OK
*rewind
btape: stored/btape.cc:581-0 Rewound "tapedrive-0"
(/dev/tape/by-id/scsi-350050763121a063c-nst)
*scan
btape: stored/btape.cc:1901-0 Starting scan at file 0
btape: stored/btape.cc:1909-0 Bad status from read -1.
ERR=stored/btape.cc:1907 read error on
/dev/tape/by-id/scsi-350050763121a063c-nst. ERR=Cannot allocate memory.
```
but btape scanblocks works fine:
```
]$ btape /dev/tape/by-id/scsi-350050763121a063c-nst
Tape block granularity is 1024 bytes.
btape: stored/butil.cc:290-0 Using device:
"/dev/tape/by-id/scsi-350050763121a063c-nst" for writing.
btape: stored/btape.cc:485-0 open device "tapedrive-0"
(/dev/tape/by-id/scsi-350050763121a063c-nst): OK
*rewind
btape: stored/btape.cc:581-0 Rewound "tapedrive-0"
(/dev/tape/by-id/scsi-350050763121a063c-nst)
*scanblocks
1 block of 203 bytes in file 0
2213 blocks of 1048576 bytes in file 0
1 block of 1048566 bytes in file 0
1072 blocks of 1048576 bytes in file 0
1 block of 1048568 bytes in file 0
```
my device configuration:
```
Device {
Name = "tapedrive-0"
DeviceType = tape
Maximum Concurrent Jobs = 1
# default:0, only required if the autoloader have multiple drives.
DriveIndex = 0
ArchiveDevice = /dev/tape/by-id/scsi-350050763121a063c-nst # Quantum
LTO-7 standalone drive
MediaType = LTO
AutoChanger = no # default: no
AutomaticMount = yes # default: no
MaximumFileSize = 40GB
MaximumBlockSize = 1048576
Maximum Spool Size = 40GB
Spool Directory = /var/lib/bareos/spool
}
```
When I read other jobs that are on the same tape (second tape of the set of
two), the restore/migration works just fine.
The question is, what is the issue here?
And how do I fix it?
Does anybody have an advice for me?
--
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 view this discussion on the web visit
https://groups.google.com/d/msgid/bareos-users/7de64c9f-6c41-4afd-ad6a-801a7bac1949n%40googlegroups.com.
