1

I've written about 1.5TB of files onto a tape with

tar -b 512 -clpMvf /dev/nst0 /path_to_stuff

Everything worked well, no error messages. Now if I try to read it back from tape with

tar -b 512 -xlpMvf /dev/nst0 /path_to_stuff

I'm getting about 10-12 read errors:

path_to_stuff/file.ext
tar: /dev/nst0: Cannot read: Input/output error

If I do a byte-by-byte comparison of the original file and the restored one, it's identical, so everything was read correctly. If I re-run, the read errors occur on different files.

The tape drive is a SAS-connected LTO-7 by IBM. Firmware was recently updated, didn't change anything.

This is what tapeinfo -f is giving me:

Product Type: Tape Drive
Vendor ID: 'IBM     '
Product ID: 'ULTRIUM-HH7     '
Revision: 'G9Q1'
Attached Changer API: No
SerialNumber: '1132T125132'
MinBlock: 1
MaxBlock: 8388608
SCSI ID: 1
SCSI LUN: 0
Ready: yes
BufferedMode: yes
Medium Type: 0x78
Density Code: 0x5c
BlockSize: 0
DataCompEnabled: yes
DataCompCapable: yes
DataDeCompEnabled: yes
CompType: 0xff
DeCompType: 0xff
BOP: yes
Block Position: 0
Partition 0 Remaining Kbytes: -1
Partition 0 Size in Kbytes: -1
ActivePartition: 0
EarlyWarningSize: 0
NumPartitions: 0
MaxPartitions: 3

Any ideas are highly appreciated!

vfxdev
  • 111
  • 2
  • 1
    Stab in the dark...could it be an issue with your multi-volume switch? – jiveturkey May 02 '17 at 15:56
  • I tried removing it the multi-volume switch, still getting the errors. Thanks anyway! – vfxdev May 02 '17 at 16:19
  • I don't know why you're hitting I/O error, but setting your blocking factor to 16384 would probably improve performance a lot. You can calculate this value by taking "MaxBlock" from tapeinfo and divide it by 512. – Andreas Rogge Feb 14 '19 at 20:46

1 Answers1

0

I have the same issue in the last few days.

The problems occurred after the last system upgrade. I watched the behavior of the system during the backup and could see that the entire memory was being used. At the same time, the reading errors occurred on the tar. During this upgrade the system swap was configured but not activated.

So I activated the swap again and started the same tar calls. Surprise! The errors did not occur anymore. Even though only a few kb were used by the swap (system memory 32G swap 16G).

Mario Goppold
  • 36
  • 4