This one time, at band camp, DK Smith wrote: >Are we sure that this answers the question that was asked? i.e. is it >the >correct answer in limited number of cases but not all cases? Is your >test case a coincidence?
Well, I'm not 100% sure. I only noticed a correlation between the output of amadmin export and my expected results, so this all hinges on the output of amadmin export being correct. A corollary of my finding is that an amadmin import only ever gets the most recent tape set imported -- I've never used amadmin import/export before so I have no idea if this is the required behaviour. >This is certainly motivating me to take a closer look at my tapes. I >thought >I could restore in much fewer tapes than what that pipeline indicates. >However, it's late and I'm tired, so I may not be grokking this. The question I'd asked is which tapes do we need to keep offsite in order to have a complete set of tapes to restore any DLE in an emergency. That's the set of tapes that contain the latest level 0, and any level 1 after that level 0, and any level 2 after that level 1, and so on, for each DLE in the network. In the best case, it's the last runtapes tapes, but that number can be pushed up due to delayed dumps, and amflush runs; also tapes may be skipped for whatever reason making a non-contiuguous list. -- [EMAIL PROTECTED] http://spacepants.org/jaq.gpg
