PORT-WRITE/PORT-DUMP questions with tape spanning
2007-07-13 12:36:01
Hi,
What happens when a DLE on a client goes to tape directly (PORT-WRITE)
because it can't fit in the holding disk and it bigger than tape
capacity? Is it possible to use tape spanning in that case?
One more: I have this DLE on a client using a tape-spanning dumptype
along with calcsize. Previously I was using excludes/includes but
yesterday I decided to have a go with spanning. I thought I'd fool
amanda by not forcing a full dump but then this situation happens:
calcsize reported a incremental of 32MB but that was way off as 11
chunks of 9GB were written to tape (8 on one tape and 3 others on the
next tape) then amanda reported 'dump to tape filed' and gave up.
Since that DLE is over 200GB and the holdingdisk use is 100GB I
suspect amanda gave because it thought that there was no more holdind
disk. amdump Right?
I can provide the entire amdump logfile and tape content is someone
is willing to enlighten me but here are the last lines before
quitting:
taper: reader-side: got label stk_40-conf3-000001 filenum 2
taper: reader-side: got label stk_40-conf3-000001 filenum 3
driver: state time 58951.903 free kps: 1025400 space: 4954528 taper: writing
idle-dumpers: 12 qlen tapeq: 0 runq: 0 roomq: 0 wakeup: 0 driver-idle:
no-dumpers
driver: interface-state time 58951.903 if default: free 1024000 if local: free
1000 if le0: free 400
driver: hdisk-state time 58951.903 hdisk 0: free 4954528 dumpers -2
driver: result time 58951.903 from taper: DONE 00-00056 stk_40-conf3-000001 3
"[sec 6827.270 kb 99903360 kps 14633.0 {wr: writers 172861 rdwait 0.000 wrwait
359.776 filemark 2.955}]"
driver: finished-cmd time 58955.920 taper wrote
gaspar:/export_raid06/koenigstiger1
driver: state time 58955.920 free kps: 1025400 space: 104857600 taper: idle
idle-dumpers: 12 qlen tapeq: 0 runq: 0
thanks,
jf
--
<° ><
|
<Prev in Thread] |
Current Thread |
[Next in Thread>
|
- PORT-WRITE/PORT-DUMP questions with tape spanning,
Jean-Francois Malouin <=
|
|
|