The following information can also be obtained from REST API on 0.20.5, right ?
Thanks On Fri, Jul 23, 2010 at 4:31 PM, Stack <[email protected]> wrote: > Read -ROOT- table to find that. It'll usually have one row only with > detail on the .META. region including info:server column which has > meta server location. > St.Ack > > > On Fri, Jul 23, 2010 at 4:07 PM, Ted Yu <[email protected]> wrote: > > For item 1, I was also looking to find out the region server name which > > hosts .META. > > > > Thanks > > > > On Fri, Jul 23, 2010 at 3:51 PM, Stack <[email protected]> wrote: > > > >> On Fri, Jul 23, 2010 at 2:40 PM, Ted Yu <[email protected]> wrote: > >> > Hi, > >> > I am looking for HBase API for getting the following information: > >> > 1. is Meta table available - I found > >> > RegionServerOperation.metaTableAvailable() but it is protected > >> > >> Thats not what you think it is Ted. > >> > >> To see if meta is available, do a Scan on it every so often. You can > >> open a Scanner -- via thrift if you are using perl, or write a bit of > >> jruby and have it return a non-zero exit code on fail -- on it as you > >> would any other. If the scan does not complete in some time then its > >> not deployed. > >> > >> > >> > 2. I found ProcessRegionStatusChange.metaRegionAvailable() which is > also > >> > protected > >> > >> Yeah, this is stuff internal to the servers; it won't get you what you > >> want. > >> > >> > 3. cluster statistics: the total number of tables, the total space > >> consumed > >> > by all the tables > >> > > >> > >> Total number of tables is something you'd have to do yourself I'd say. > >> Given the above meta scan, look at the row names and when their > >> prefix changes -- their prefix is the table name the region belongs > >> too, increment your table count. > >> > >> To see how much space is being used by hbase, thats ./bin/hadoop fs > >> -dus /HBASE.ROOTDIR parsing whatever it returns. > >> > >> St.Ack > >> > >> > The goal is to call the above APIs from Perl library our Ops has > >> developed > >> > to facilitate cluster management. > >> > > >> > Thanks > >> > > >> > > >
