I take it that ycsb is working for you now? On Wed, Feb 16, 2011 at 4:27 PM, Ryan Rawson <[email protected]> wrote:
> +1, been running this with HDFS-347 and a VerifyingWorkload and no errors! > > On Wed, Feb 16, 2011 at 4:02 PM, Todd Lipcon <[email protected]> wrote: > > +1. On a 5 node test cluster, loaded 100M rows (~170G on disk) using > YCSB, > > did some mixed workload (increment, scan, read, put). No GC pauses. > Killed a > > server while under load, recovered correctly. > > > > Along the way I accidentally deleted my ZK jar while the cluster was > > running, which left a couple regions unassigned. hbck -fix fixed them. > > > > Also, the new feature that checks for out-of-date hbase-default.xml > detected > > that I had left one lying around on the classpath from a previous > version, > > so that works! :) > > > > -Todd > > > > On Wed, Feb 16, 2011 at 11:03 AM, Stack <[email protected]> wrote: > > > >> +1 > >> > >> Downloaded, reviewed doc., and checked it runs locally. > >> > >> Put it up on cluster, ran some YCSBs -- I deployed a snapshot of RC1 > >> to apache maven snapshot repo with name 0.90.1-20110215.213202-4 if > >> anyone needs it -- and all seemed to run fine in basic write and > >> read/write scenarios. Review of logs turned up no exceptions. > >> > >> I ran rowcounter to make sure MR works as documented. > >> > >> St.Ack > >> > >> > >> On Mon, Feb 14, 2011 at 5:37 PM, Todd Lipcon <[email protected]> wrote: > >> > The HBase team is pleased to announce the second [zero-indexed] > release > >> > candidate for HBase 0.90.1, available for download at the following > URL: > >> > http://people.apache.org/~todd/hbase-0< > >> http://people.apache.org/~todd/hbase-0.90.1.rc0/> > >> > - HBASE-3515 [replication] ReplicationSource can miss a log after > RS > >> > comes > >> > <http://people.apache.org/~todd/hbase-0.90.1.rc0/> > >> > - out of GC > >> > .90.1.rc1/ <http://people.apache.org/~todd/hbase-0.90.1.rc0/> > >> > > >> > This release candidate addresses a couple of bugs since the last > release > >> > candidate: > >> > HBASE-3524 NPE from CompactionChecker > >> > HBASE-3525 mvn assembly is over-filling the hbase lib dir > >> > HBASE-3531 When under global memstore pressure, dont try to flush > >> > unflushable regions. > >> > > >> > > >> > The following is the md5sum: > >> > dbf090fcb12a47c0b6c713c3caa7b1fd hbase-0.90.1.tar.gz > >> > and it corresponds to the 0.90.1RC1 tag at r1070711. > >> > > >> > <http://people.apache.org/~todd/hbase-0.90.1.rc0/>HBase 0.90.1 is a > >> minor > >> > release containing important bug fixes and low-risk improvements since > >> > HBase 0.90.0. The following is a complete list of tickets resolved for > >> this > >> > minor release: > >> > > >> > > >> > https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12310753&version=12315548 > >> > > >> > As a minor release, no migration procedures are necessary to upgrade > from > >> > 0.90.0. Additionally, no public facing APIs should differ in any way. > If > >> you > >> > are upgrading from a previous major release, some caveats may apply - > >> please > >> > see the relevant documentation for details. All users of 0.90.0 are > >> advised > >> > to upgrade at their earliest convenience. > >> > > >> > Please vote by Thurs 2/17 whether we should release this artifact as > >> HBase > >> > 0.90.1. PMC votes are binding, non-PMC votes are as always very much > >> > appreciated! > >> > > >> > Thanks > >> > The HBase Team > >> > > >> > > >> > -- > >> > Todd Lipcon > >> > Software Engineer, Cloudera > >> > > >> > > > > > > > > -- > > Todd Lipcon > > Software Engineer, Cloudera > > >
