[Summary] ufsdump write error on tape 9, restart?

Amazing (the list, that is).

Thanks to Bill Voight, Jay Lessert, Rich Kulawiec, Sysadmin@astro.su.se.

Turns out Rich wrote a significant amount of the original code that then 
became ufsdump. His explanations was that the ufsdump spawns a child for 
the new tape, giving it instructions on how to proceed. If I answer the 
child with a "yes" (start over), it will return to the parent who will 
see that it wasn't done and spawn another child with the same instructions.

Therefore, a "yes" answer will cause tape number 9 to be rewritten, and 
will not go all the way back to the beginning with tape 1. A "no" answer 
would have aborted the backup. The results:

  DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")
  DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no") yes
  DUMP: 88.80% done, finished in 3:19
  DUMP: Change Volumes: Mount volume `#9' on 
`images.library.umass.edu:/dev/rmt/0n'
  DUMP: NEEDS ATTENTION: Is the new volume (#9) mounted on 
`images.library.umass.edu:/dev/rmt/0n' and ready to go?: ("yes" or "no") yes
  DUMP: Volume 9 begins with blocks from inode 136184
  DUMP: 89.34% done, finished in 3:09

(I used the cleaning tape after manually ejecting the first tape #9, and 
then put in a new tape #9)


--------------------

I was also told by more than one that I shouldn't beat up on my internal 
drive so much. Get an autoloader. I agree.

One of my summer projects is to put the Veritas client on this server 
and send the D1000 backup over to an LTO autoloader.

---------------

Chris Hoogendyk

-
   O__  ---- Network Specialist & Unix Systems Administrator
  c/ /'_ --- Library Information Systems & Technology Services
 (*) \(*) -- W.E.B. Du Bois Library
~~~~~~~~~~ - University of Massachusetts, Amherst 

<choogend@library.umass.edu>

--------------- 


Origianl Question:

Sorry to bother the list with such a simple question that I can actually 
answer myself by just doing it. However, I'd rather know the answer than 
experiment. Notice that it is tape 9. In other words, it has 
successfully gotten through 8 tapes (DDS/3, Sony Premium, brand new), is 
almost done, and gives an error on what would be the last tape. It took 
three days to get to that point (partly because I don't have 24 hour 
operations staff).

It is an E250 with Solaris 8 and the standard DDS/3 internal drive that 
comes with the E250. Backing up scanned image files from a D1000.

Question: If I say "yes", does that mean start the ENTIRE dump over? or 
just that tape? Can I eject the tape, put in a fresh one (or clean the 
drive and then put in a fresh one), and then have it begin again with 
the 9th tape?

Or, is the answer "no" to mean "no, don't restart the whole thing, just 
do the 9th tape again?" Or will "no" cause the whole thing to abort?

I checked the admin manuals, and I also looked at the database of error 
messages on docs.sun. Couldn't find an explanation.


 DUMP: NEEDS ATTENTION: Is the new volume (#9) mounted on 
`images.library.umass.edu:/dev/rmt/0n' and ready to go?: ("yes" or "no") yes
 DUMP: Volume 9 begins with blocks from inode 136184
 DUMP: 88.80% done, finished in 3:19
 DUMP: Write error 17 feet into tape 9
 DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")
 DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")



TIA
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers