On Thu, Feb 19, 2009 at 4:26 PM, stan <[email protected]> wrote: > Notice that at the end, it looks for the next tape label "DailyDump05" but > it thinks it should be slot 4, vtape 4 (DailyDump05 is actually slot 6, vtape > 6)! > It looks like it is decrementing the count by 1, instead of incrementing & > it must be confused about the current slot when it starts.
It's getting a failure writing to the first tape, and then seeking sequentially for subsequent tapes. It's the initial failure that's problematic. It's occurring in the eight seconds here: (taper writing tape labels to the two RAIT child devices:) 1235075063.161127: taper: blocksize = 32768 ... (taper tells driver it's starting a new tape) 1235075066.041323: taper: putresult: 18 NEW-TAPE (taper asks driver for permission to use another (second) tape, due to an error with the first) 1235075071.400399: taper: putresult: 26 REQUEST-NEW-TAPE The taper never began writing the dump file, or we would see type=DUMPFILE headers in the logfiles. Can you attach the amdump logfile (amdump.*) and the trace logfile (log.*.1) from this same run? Taper sends some of its messages to each of those logfiles. Dustin -- Storage Software Engineer http://www.zmanda.com
