Hi Folks,
since some time I am using Amanda and everything worked fine but some day,
without me having changed anything, amdump started to fail and is still
failing. The problem is reproduceable on a similar systemconfiguration with
the same tape-drive. I have attached some lines of the log file produced by
amanda.
To the system configurations:
First:
- Kernel 2.2.14 on an 200 MHz Pentium
- 64 Mb RAM
- Adpatec with AIC7881U
- Tandberg SLR 50 Streamer (25Gb uncompressed)
- 30 Gb holding disk (IDE [IBM-DLTA 307030])
- amanda 2.4.2p1
- mt version 0.5b
Second:
- Kernel 2.2.19 an a 800 MHz P III
- 512 Mb RAM
- ICP-Vortex 6118RS Raid Controller
- Tandberg SLR 50 Streamer (25Gb uncompressed)
- 10 Gb holding disk (SCSI)
- amanda 2.4.2p1
- mt version 0.5b
The main problem is that I permanently get I/O Error while writng Filemark
or while closing the tape. But this happens only during night while amdump
ist running. Amflush causes no problems.
Now to my questions:
- Does anyone have an idea what could cause the problem?
- Does anyone have a proper tapetype definition for this streamer?
Greetings
Nicki Messerschmidt
P.S.: The Logfile parts (taken from the mails I got):
*** A TAPE ERROR OCCURRED: [[writing file: Input/output error]].
taper: FATAL syncpipe_get: w: unexpected EOF
planner: Last full dump of blasebalg.touristicguide.com:sda5 on tape
bb-backup_002 overwritten in 4 runs.
planner: Incremental of pluto.link-m.de:sda5 bumped to level 2.
planner: Incremental of merkur.link-m.de:sda6 bumped to level 3.
taper: tape bb-backup_008 kb 352352 fm 7 writing file: Input/output error
driver: going into degraded mode because of tape error.
