-1
The log spew doesn't happen always but if it did, as per Andrew, could fill
a partition and mess up a cluster.
HBASE-15218 though an oldie, is a baddie.
The combo sinks the RC for me.
On up side, I've been running cluster tests and 1.2.0 is looking good.
St.Ack
On Fri, Jan 29, 2016 at
This vote fails with a single +1 and no other votes.
Thanks to everyone who evaluated the release candidate!
-Sean
On Fri, Jan 29, 2016 at 9:29 AM, Sean Busbey wrote:
> Hi Folks!
>
> I'm pleased to announce the second release candidate for HBase 1.2.0.
>
> Artifacts are
Sean
If you are going to make a new RC, I would like to get HBASE-15218 fixed in
it. This is not particularly broke in 1.2.0 but from 1.0.2 time.
-Anoop-
On Fri, Feb 5, 2016 at 2:16 AM, Andrew Purtell
wrote:
> Logging lines at ~milliseconds frequency could certainly
Friendly reminder that this vote closes tomorrow.
Stack, do you have a better idea about the severity of HBASE-15207 and
wether the root cause is new to 1.2.0?
-Sean
On Tue, Feb 2, 2016 at 1:37 PM, Stack wrote:
> It looks like balancer got stuck after Master joined running
Meant to report back here. I've been running loadings with monkeys trying
to repro 'hung balancer'. Now I am of the opinion that there is no 'hung
balancer', just balancer spew. The spew came in on HBASE-13376 which would
be new to 1.2. Want to wait on at least this big cycle to finish before
Logging lines at ~milliseconds frequency could certainly lead to a production
problem. For example log files filling a volume rendering a master node out of
service. If possible can we fix it and spin a new RC?
> On Feb 4, 2016, at 12:18 PM, Stack wrote:
>
> Meant to report
It looks like balancer got stuck after Master joined running cluster (it
had been killed by Monkey). We then log at a rate of about 10 lines per
millisecond. HBASE-15207
St.Ack
On Tue, Feb 2, 2016 at 9:51 AM, Stack wrote:
> I've been running some cluster loadings on the RC.
I've been running some cluster loadings on the RC. Last night my logs
filled with this (10x256MB log files):
2016-02-01 11:25:26,958 DEBUG
[B.defaultRpcServer.handler=9,queue=0,port=16000]
balancer.BaseLoadBalancer: Lowest locality region server with non zero
regions is
Hi Folks!
I'm pleased to announce the second release candidate for HBase 1.2.0.
Artifacts are available here:
https://dist.apache.org/repos/dist/dev/hbase/hbase-1.2.0RC1/
As of this vote, the relevant md5 hashes are:
a338ca93cd4c495f03bcff2d457222ef hbase-1.2.0-bin.tar.gz
That closing date should be 4:00PM UTC
on Friday, Feb 5th, 2016 and not 2015. The link to the "find this in
your local timezone" has the correct date/time.
On Fri, Jan 29, 2016 at 9:29 AM, Sean Busbey wrote:
> Hi Folks!
>
> I'm pleased to announce the second release candidate
+1
Ran test suite using Java 1.7.0_67
Exercised basic shell commands
[INFO] Apache HBase - Shaded - Client . SUCCESS [
0.493 s]
[INFO] Apache HBase - Shaded - Server . SUCCESS [
0.867 s]
[INFO]
Compilation didn't pass on the expanded source tar ball:
[ERROR]
/home/hbase/hbase-1.2.0/hbase-server/src/main/java/org/apache/hadoop/hbase/master/procedure/MasterProcedureQueue.java:[60,7]
error: MasterProcedureQueue is not abstract and does not override
abstract method poll() in
Oops - I should have removed the old directory before untarring.
Running tests now.
Will report back.
On Fri, Jan 29, 2016 at 10:49 AM, Matteo Bertozzi
wrote:
> that file MasterProcedureQueue is no longer in 1.2.
> maybe something from the old RC was left in the dir
>
that file MasterProcedureQueue is no longer in 1.2.
maybe something from the old RC was left in the dir
Matteo
On Fri, Jan 29, 2016 at 10:39 AM, Ted Yu wrote:
> Compilation didn't pass on the expanded source tar ball:
>
> [ERROR]
>
>
14 matches
Mail list logo