You have it Norbert. The timestamp is actually time of region creation (which yes would correlate with region split time since this is when the daughter regions are created).
St.Ack On Mon, Jul 18, 2011 at 8:40 PM, Norbert Burger <[email protected]> wrote: > Thanks, Stack. I realize now that I most likely confused a timestamp that > happens into exist in our table's row key with one that (possibly) refers to > the region split time. Is it correct, then, to say that rowkeys from .META. > have the following format? > > tableName,regionStartRow,timestampRegionSplit.someMD5Hash > > Norbert > > > > > > > > > > > > > > sure I confused I confused a timestamp which happens to exist in the > table's rowkeys in question, with > > On Mon, Jul 18, 2011 at 11:08 PM, Stack <[email protected]> wrote: > >> Please do a little text diagram that shows the "timestamps" to which >> you are referring in a sample .META. row (Perhaps you take the md5 >> hash on the tail of the region name as a 'timestamp'?) >> >> St.Ack >> >> On Mon, Jul 18, 2011 at 3:28 PM, Norbert Burger >> <[email protected]> wrote: >> > Hey folks -- >> > >> > For all rowkeys in .META. except for the last region of a table, I see a >> > couple of different timestamps. Could someone confirm which one is >> > region-splitted-at timestamp, and what the logical meaning of the other >> > timestamp is? >> > >> > Thanks, >> > Norbert >> > >> >
