Why is Zookeeper Listener not abstract ?

It looks like arg.apache.hadoop.hbase.zookeeper.ZooKeeperListener could be better described as an abstract class. is there any specific reason why its not made abstract ? Cheers, Akash A

Re: Why is Zookeeper Listener not abstract ?

A ? > > > > On Tue, Jun 21, 2011 at 7:21 AM, Akash Ashok > > wrote: > > > > > It looks like arg.apache.hadoop.hbase.zookeeper.ZooKeeperListener could > > be > > > better described as an abstract class. is there any specific reason why > > its > > > not made abstract ? > > > > > > Cheers, > > > Akash A > > > > > >

Re: Why is Zookeeper Listener not abstract ?

Sure. Thanks :) On Tue, Jun 21, 2011 at 10:20 PM, Ted Yu wrote: > I am not sure putting exclamation mark in JIRA title is a good idea. > So I removed it. > > Pardon me. > > On Tue, Jun 21, 2011 at 9:48 AM, Akash Ashok > wrote: > > > Hi, > >

Inappropriate conditionLOG.isDebugEnabled() in HregionServer.java

stop-hbase.sh never stopped once deployed the 0.91 trunk jar onto my hbase setup. Figured out that the Meta regions were never closed and hbase would stop only when logging in debug mode is enabled. Here is the code snippet from HResionServer.java on HBaseTrunk: else if (this.stopping && LOG.isDe

Re: Inappropriate conditionLOG.isDebugEnabled() in HregionServer.java

ink in isOnlyMetaRegionsRemaining(), onlyMetaRegionsRemaining > should be initialized to true so that the cleanup at line 638 can proceed > if > this.onlineRegions is empty. > > Mind filing a bug ? > > Thanks > > On Fri, Jun 24, 2011 at 5:52 AM, Akash Ashok > wrote: &g

Re: Inappropriate conditionLOG.isDebugEnabled() in HregionServer.java

On Fri, Jun 24, 2011 at 7:52 PM, Ted Yu wrote: > I think the if condition below can be kept. It is an optimization because > isOnlyMetaRegionsRemaining() would loop through > this.onlineRegions.entrySet() > > On Fri, Jun 24, 2011 at 7:08 AM, Akash Ashok > wrote: > > &

JVMClusterUtil.startup method refactor

Hi, I was having a look at the org.apache.hadoop.hbase.util.JVMClusterUtil. The startup method has the following signature public static String startup(final List masters, final List regionservers) The return type is a string which return the ActiveMaster Information in the string format whi

Re: JVMClusterUtil.startup method refactor

On Fri, Jul 15, 2011 at 10:52 PM, Stack wrote: > On Fri, Jul 15, 2011 at 10:20 AM, Akash Ashok > wrote: > > Hi, > >I was having a look at the > org.apache.hadoop.hbase.util.JVMClusterUtil. > > The startup method has the following signature > > > >

Re: JVMClusterUtil.startup method refactor

I decided against it as the change i was thinking about broke a lot of test cases. So I'm gonna leave it as it is. :) On Sat, Jul 16, 2011 at 2:10 PM, Akash Ashok wrote: > > > On Fri, Jul 15, 2011 at 10:52 PM, Stack wrote: > >> On Fri, Jul 15, 2011 at 10:20 AM, Akash A

heapSize() implementation of KeyValue

Hi, I was going thru the heapSize() method in the class KeyValue and i couldn't seem to understand a few things which are in bold below private byte [] bytes = null; private int offset = 0; private int length = 0; private int keyLength = 0; // the row cached private byte [] rowC

Re: heapSize() implementation of KeyValue

array (hence the references), > then we are taking account of the overhead of the 'bytes' array > itself. > > And I see 3 integers pasted in, so things are looking good to me > > On Sun, Jul 31, 2011 at 10:01 PM, Akash Ashok > wrote: > > Hi, > > I was g

Re: heapSize() implementation of KeyValue

Sorry I think thats the length of the Array being taken care of not the size of the int itself. My bad i confused it with size of the "int" itself. I think that explains it. Thanks a lot :) On Mon, Aug 1, 2011 at 10:51 AM, Akash Ashok wrote: > Three integers are pasted in but

HBase book possible typo

Hi, In the HBase book Architecture -> Storage -> Overview Its quoted that " The HRegionServer opens the region and creates a corresponding HRegion object. When the HRegion is opened it sets up a Storeinstance for each * HColumnFamily* for every table as defined by the user beforehand." S

Re: HBase book possible typo

9:46 AM, "Doug Meil" wrote: > > > > >Thanks. I'll follow-up on it. > > > > > > > > > > > >On 8/22/11 2:04 AM, "Akash Ashok" wrote: > > > >>Hi, > >> In the HBase book Architecture -> Storage -> Overvie

Why is there only one instance of HLog.?

Hi, I was just wondering why is there only one instance of HLog? Wouldn't this incur the overhead of synchronization when more than one region tries to write to ? Why dnt we have an HLog per HRegion so that they Hregion's writes are independent of each other? Plz do let me know if I am missing

Re: Why is there only one instance of HLog.?

plementation on each GFS server, these writes could cause a large number of disk seeks to write to the different physical log files." Cheers, Akash A On Tue, Aug 23, 2011 at 3:18 AM, Akash Ashok wrote: > Hi, > I was just wondering why is there only one instance of HLog? Wouldn'

Why doesn't HConnectionManager implement double checked locking ?

Hi, I have pasted the code below for HConnectionManager.getConnection public static HConnection getConnection(Configuration conf) throws ZooKeeperConnectionException { HConnectionKey connectionKey = new HConnectionKey(conf); synchronized (HBASE_INSTANCES) { HConnectionImplem

Re: Why doesn't HConnectionManager implement double checked locking ?

.org/wiki/Double-checked_locking) -- > help here? What you thinking? We can't really do putIfAbsent w/ a > connection instance? > > Good on you Akash, > St.Ack > > On Tue, Aug 23, 2011 at 11:45 AM, Akash Ashok > wrote: > > Hi, > > I have p

Re: Why doesn't HConnectionManager implement double checked locking ?

t.Ack > > On Thu, Aug 25, 2011 at 4:55 AM, Akash Ashok > wrote: > > I was just curious as to whether this was a hotspot or not. I mean as > Ryan > > said if its not a part of high perf code path then its ok. > > > > Also now I'm wondering why can't we do

Running UnitTests before submitting a patch

Hi, I was just wondering as to how developers go about running unit tests before they wanna submit a patch. I am not asking about testing the functionality of the patches, but about the whole suite of regression test cases. Firstly I end up running the whole suite without a the patch and second

Re: Running UnitTests before submitting a patch

Hi, FAIL SAFE PLUGIN: FailSafe Plugin sounds like a great idea. My understanding is that we startup and kill the hdfs, zookeeper and HBase cluster instances for every test suite ( Please correct me if I am wrong here ) . . With FailSafe plugin we would be able to achieve the following things: 1.

Re: Running UnitTests before submitting a patch

Hi All, Just a suggestion about the running tests. We should also enable re-running or retrying the failed or skipped test cases just to make sure the result wasn't just an erroneous one. Considering the discussion started by Ted in mail chain below this would result in better outcome thus

Re: Running UnitTests before submitting a patch

king about. Added a new ticket > (HBASE-4480<https://issues.apache.org/jira/browse/HBASE-4480>) > on this issue. > > -Jesse Yates > > On Sat, Sep 24, 2011 at 12:11 PM, Akash Ashok >wrote: > > > Hi All, > >Just a suggestion about the running tests. W

Re: Running UnitTests before submitting a patch

ep 25, 2011 at 1:50 PM, Akash Ashok wrote: > Jesse, > > We had a similar situation in our project but we figured Maven itself > can't help us out with this. But JUnit can. This might mean an annotation > might be required to be added to all the test cases. > > We h

Re: Running UnitTests before submitting a patch

T > > Cheers, > Akash A > > > On Sun, Sep 25, 2011 at 1:50 PM, Akash Ashok wrote: > >> Jesse, >> >> We had a similar situation in our project but we figured Maven itself >> can't help us out with this. But JUnit can. This might mean an annotation

Re: Better Javadoc HTableDescriptor

While I was doing this I found out that there is a method in HTableDescriptor called isMetaTable() which basically checks if the table is a catalog table and the name kind of misleading. I am planning to rename this method to isCatalogTable(). 1. Is it ok to do this renaming considering that this

Multiple WALs

Hey, I've see that setting writeToWAL(false) boosts up the writes like crazy. I was just thinking having MuiltipleWAL on HBase. I understand that this is a consideration in BigTable paper that a WAL per region is not used because it might result in a lot of disk seeks when there are large number of

Re: Better Javadoc HTableDescriptor

hat may happen on a major upgrade, but it had > better be for a very good, explainable, reason that the community is > behind. > > > A fourth option is to add Javadoc to the explaining what this method does > and not add any new methods. > > > On 10/1/11 8:02 AM, "Akas

Re: Better Javadoc HTableDescriptor

s that causes compilation failures > upon > > > upgrade. Once in a while that may happen on a major upgrade, but it > had > > > better be for a very good, explainable, reason that the community is > > > behind. > > > > > > > > > A fourth opt

Re: Multiple WALs

t; > However, irrespective of whether I have one wal or two, I have > > seeing > > > the > > > >> > same throughput. I have experimented with an HDFS setting that > > allows > > > >> > writing/sync to multiple replicas in parallel, and tha

Re: Multiple WALs

Uh Oh! Its showing the reporter of the problem as me but it wasn't me in reality :) I am not able to modify it. Please feel free to change it :) Cheers, Akash A On Sun, Oct 2, 2011 at 10:38 AM, Akash Ashok wrote: > I've opened up a JIRA for this > https://issues.apache.org/j

Re: mvn error?

This is because of the surefire reports plugin version. Works well with 2.7.2 but not with 2.8 and above Similar discussion can be found in http://stackoverflow.com/questions/5339373/maven-surefire-report-plugin-2-8-error-org-apache-maven-doxia-siterenderer-sink Cheers, Akash A On Mon, Oct 3, 201

Re: Build failed in Jenkins: HBase-TRUNK #2287

Hey Doug, Actually this is an issue with maven-site-plugin 2.1 not with the acid-semantics.xml. Just figured out that moving maven-site-plugin back to 2.0.1 and downgrading surefire-reports plugin to 2.7.2 does the trick. Cheers, Akash A On Tue, Oct 4, 2011 at 5:56 AM, Doug Meil wrote: > > This

Re: Build failed in Jenkins: HBase-TRUNK #2287

Sure Stack. Shall submit it out in sometime. On Tue, Oct 4, 2011 at 9:39 AM, Stack wrote: > Mind making a patch Akashi? > > > > On Oct 3, 2011, at 19:56, Akash Ashok wrote: > > > Hey Doug, > > Actually this is an issue with maven-site-plugin 2.1 not with th

Re: Build failed in Jenkins: HBase-TRUNK #2287

Hi Stack, I've submitted the patch. Its diff'(ed) on HEAD of trunk. Cheers, Akash A On Tue, Oct 4, 2011 at 6:09 PM, Akash Ashok wrote: > Sure Stack. Shall submit it out in sometime. > > > On Tue, Oct 4, 2011 at 9:39 AM, Stack wrote: > >> Mind making a patch Aka

Re: Jenkins build is back to normal : HBase-TRUNK #2304

Oh cool. Build is back to normal. Could someone tell me what the issue was. Why was it failing even though there were no failures ? On Sat, Oct 8, 2011 at 4:45 AM, Apache Jenkins Server < jenk...@builds.apache.org> wrote: > See > > >

Re: Jenkins build is back to normal : HBase-TRUNK #2304

> > execute goal org.apache.maven.plugins:maven-surefire-plugin:2.9:test > > > (default-test) on project hbase: Failure or timeout > > > > > > Unluckily we don't have access to the build machine. > > > > > > On Fri, Oct 7, 2011 at 10:14 PM, Aka

HTableDescriptor Values Map Usage

Hi, Looking the the values map that is currently being used in HTableDescriptor its currently being used for Storing 1. if the table is -ROOT- 2. is table .META. 3. RegionSplitPolicy Classname 4. CoProcessors attached to the current table. Out of which only co processors stored in the map makes s

Re: HTableDescriptor Values Map Usage

My bad root and meta are a part of instance variables. If that is the case can someone please tell me why is it first being stored in the values map? Cheers, Akash A On Sun, Oct 9, 2011 at 6:58 AM, Akash Ashok wrote: > Hi, > Looking the the values map that is currently being u

Re: Better Javadoc HTableDescriptor

(); > } > > public boolean isCatalogTable() { >return isMetaRegion() && !isRootRegion(); > } > > Thanks, > Shumin > > On Sat, Oct 1, 2011 at 1:16 PM, Akash Ashok > wrote: > > > Ok Thanks. I shall just add the javadoc. I've already o

HBase Trunk Hung

Hi, HBase Trunk is hung and its not progressing and there have be no runs from October 13th. Any1 has any idea about this ? Cheers, Akash A

Re: HBase Trunk Hung

https://builds.apache.org/job/HBase-TRUNK/ On Sun, Oct 16, 2011 at 8:44 AM, Akash Ashok wrote: > Hi, > HBase Trunk is hung and its not progressing and there have be no runs > from October 13th. Any1 has any idea about this ? > > Cheers, > Akash A >

Re: HBase Trunk Hung

Oh ok. Thanks Ted. Cheers, Akash A On Sun, Oct 16, 2011 at 8:52 AM, Ted Yu wrote: > See https://issues.apache.org/jira/browse/INFRA-4033 > > I notified bui...@apache.org > > No response so far. > It has been hanging for almost 3 days. > > On Sat, Oct 15, 2011 at 8:15

Re: HBase Trunk Hung

Cheers > > On Sat, Oct 15, 2011 at 8:55 PM, Akash Ashok > wrote: > > > Oh ok. Thanks Ted. > > > > Cheers, > > Akash A > > > > On Sun, Oct 16, 2011 at 8:52 AM, Ted Yu wrote: > > > > > See https://issues.apache.org/jira/browse/IN

HBase-4486 ( Improved Javadoc for HTableDescriptor ) ready for review

Hi HBase-4486 ( Improved Javadoc for HTableDescriptor ) ready for review. :) https://issues.apache.org/jira/browse/HBASE-4486 Thanks, Akash A

Re: HBase Trunk Hung

letes the build after 3 days or so :) Cheers, Akash A On Thu, Oct 20, 2011 at 4:35 AM, Ted Yu wrote: > Now it is HBase-0.92 build 74, hanging for over a day. > > On Sun, Oct 16, 2011 at 8:01 AM, Akash Ashok > wrote: > > > I doubt Jenkins has any such thing. I think so

Re: HBase Trunk Hung

it can be configured to 300 or 400 mins that would be great. > > Sent from my iPhone > > On Oct 20, 2011, at 10:23, Akash Ashok wrote: > > > Yup its still Hung. Can't one of the committers ask for the permission to > > delete a build ? Sounds like a easier fix than

Re: [unit tests] small tests vs. others

Ooo +1 on this. This sounds Just wonderful. :) Cheers, Akash A On Fri, Oct 21, 2011 at 3:34 AM, Andrew Purtell wrote: > +1, this sounds great. > > In 0.20 I had the REST (aka Stargate at that time) tests running all in > single JVM but in the port to 0.90 had to be updated for the current > sur

Re: suggestion for smoother code review process

On Fri, Oct 21, 2011 at 3:35 AM, Todd Lipcon wrote: > Hey Ted, > > I agree the formatting of the reviewboard comments back onto JIRA > could be improved. I wrote the original script that does it - it's > some nasty procmail and python. > Hey Todd, I would like to work on this. Also Is it nasty pr

Re: SILT - nice keyvalue store paper

I was running some similar tests and came across a surprising finding. I compared reads and write on ConcurrentSkipListMap ( which the memstore uses) and synchronized TreeMap ( Which was literally treemap synchronized). Executed concurrent reads, writes and deletes on both of them. Surprisingly syn

Re: SILT - nice keyvalue store paper

Values corrected. there was a mistake in the number of records > And here are the details about the test run. > 100 Threads each fetching 10,000 records > 100 threads each adding 10,000 records. > 100 threads each deletin 10,000 records > ( Reads, Writes and deletes simultaneously ) > > Cheers, >

Re: SILT - nice keyvalue store paper

en't made their way to apache. > > In the end, after our optimizations, the real world workload Dhruba and I > were working with was not all in-memory so the bottleneck in production > became the random reads (so increasing the block cache hit ratio is the > focus) rather th

Re: CSLM performance Was: SILT - nice keyvalue store paper

y by reducing the number of operations and > > > increasing efficiency on the CSLM. Doing things like always taking a > > tailMap > > > and working from that instead of starting at the root node, using an > > iterator() > > > and taking advantage of t

Fwd: warning from dev@hbase.apache.org

Any Idea ? Got a warning for no reason :) Cheers, Akash A -- Forwarded message -- From: Date: Wed, Nov 16, 2011 at 3:40 AM Subject: warning from dev@hbase.apache.org To: thehellma...@gmail.com Hi! This is the ezmlm program. I'm managing the dev@hbase.apache.org mailing list. I'

Contacting a wrong regions

Hi All, We currently have an issue in our HBase cluster that 2 of our regions on a regions server have gone down. We can't enable or disable the regions. Are there any tools to get the data out of it from HDFS files ? We are ok with recreating the indices only for the keys in that region. Cheers,

Region Corrupt.

Hi All, We currently have an issue in our HBase cluster that 2 of our regions on a regions server have gone down. We can't enable or disable the regions. Are there any tools to get the data out of it from HDFS files ? We are ok with recreating the indices only for the keys in that region. Cheers,

Re: Contacting a wrong regions

Sorry for the wrong subject. I Changed it and sent another mail. On Fri, Dec 23, 2011 at 12:52 PM, Akash Ashok wrote: > Hi All, > We currently have an issue in our HBase cluster that 2 of our regions on a > regions server have gone down. We can't enable or disable the regions. A

Re: Region Corrupt.

Also due to some reason one of my regions is corrupt. I want to change the ENDKEY of that region in the META table. Is there any way to do this. If not can I file a JIRA for this. Would be a good feature to have admin for regions. Cheers, Akash A On Fri, Dec 23, 2011 at 12:54 PM, Akash Ashok

Re: Contacting a wrong regions

, Stack wrote: > On Thu, Dec 22, 2011 at 11:22 PM, Akash Ashok > wrote: > > Hi All, > > We currently have an issue in our HBase cluster that 2 of our regions on > a > > regions server have gone down. We can't enable or disable the regions. > > Hey Akash: >

Re: Contacting a wrong regions

Also we have a lot of holes in .META. Are there tools for 0.20.6 which we can use to patch them ? Cheers, Akash A On Wed, Dec 28, 2011 at 1:44 PM, Akash Ashok wrote: > Hi Stack, > hbck is not available on our cluster. We are running on age old 0.20.6 and > I presume hbck was made

Re: build time shortened to 1/4 of the original

This is really awesome. Thanks a lot N Keywal. Cheers, Akash A On Mon, Jan 2, 2012 at 10:49 AM, Ted Yu wrote: > Hi, > Thanks to N Keywal's wonderful work in HBASE-5064, both TRUNK builds and > PreCommit build now take only 29 minutes to complete (assuming reasonable > load on Jenkins build mach

Re: Please welcome Jon Hsieh as an HBase committer

Congrats Jon. :) Cheers, Akash A On Tue, Jan 10, 2012 at 6:31 AM, Stack wrote: > Good on you Jon. > St.Ack > > On Mon, Jan 9, 2012 at 2:18 PM, Todd Lipcon wrote: > > I'm happy to announce that the HBase PMC has elected Jonathan Hsieh as > > a new HBase committer and PMC member. Jon has done gr

Addition of an extra method to HRegionServer consensus

Hi All, There's a discussion going on HBase-4224 about adding a new method to the HRegionServer. Just to give a brief. The patch to HBase-4224 has 2 methods in HRegionServers 1. flushRegions(List regions) - which flushes a list of regions 2. flushAllRegions() - flushes all the regions of the reg

Re: Changing our logo purple to International Orange (Engineering)

+ 1 . This orange is way more appealing On Thu, Jan 12, 2012 at 5:17 AM, Jonathan Hsieh wrote: > +1. > > Can I get an "old school" purple one for my bike shed? :) > > Jon. > > On Wed, Jan 11, 2012 at 3:44 PM, Roman Shaposhnik wrote: > > > On Wed, Jan 11, 2012 at 8:42 AM, Stack wrote: > > > I