I that's the bug where disabling a table that was just split and still
has the parent entries in .META. puts it into a weird state. We can do
some deeper debugging to help you getting that table back but I
strongly suggest that you upgrade to 0.90.1 where the issue is fixed.
There's no backport
An hour to disable? That doesn't sound right at all :)
I would approach this problem like I generally do with HBase issue,
first check the master log for any weirdness regarding my problem (in
this case, grep for the table name).
Then I would look the region server log(s) of the nodes that were
I see the master is waiting and I see the exceptions but there's no
context and the timestamps don't match (what happened in the region
server at the time the second master took over), can you explain
exactly what was done to get in that state? Also could you please tell
use which hbase version
We use the HBase+Hive integration here for ad-hoc queries, I don't
understand the data duplication you're talking about... when you
create an external table you can directly query your existing tables.
We run with the latest patch posted in HIVE-1634 since we have a lot
of binary values and I made
On Fri, Feb 25, 2011 at 2:49 PM, Peter Haidinyak phaidin...@local.com wrote:
Sorry to jump in here but does HBase use Map/Reduce under the covers? I was
under the impression that HBase used the DFS of Hadoop but not Map/Reduce.
You're right, and I don't see anything that contradicts that in
...@gmail.com [mailto:jdcry...@gmail.com] On Behalf Of Jean-Daniel
Cryans
Sent: Friday, February 25, 2011 2:52 PM
To: user@hbase.apache.org
Subject: Re: Ad-hoc reports against HBase - any way? any tools?
On Fri, Feb 25, 2011 at 2:49 PM, Peter Haidinyak phaidin...@local.com wrote:
Sorry to jump
Exactly.
J-D
On Thu, Feb 24, 2011 at 2:45 PM, Nanheng Wu nanhen...@gmail.com wrote:
Sorry for trying to bring this topic back again guys, so currently in
0.20.6 is there's no way to drop a table without large amount of
flushing?
On Tue, Feb 22, 2011 at 3:04 PM, Jean-Daniel Cryans jdcry
and
then remove the regions from the meta table? Is hbase gonna get messed
up when it later needs to flush those regions out of the memstore?
Thanks!
On Thursday, February 24, 2011, Jean-Daniel Cryans jdcry...@apache.org
wrote:
Exactly.
J-D
On Thu, Feb 24, 2011 at 2:45 PM, Nanheng Wu
See http://hbase.apache.org/upgrading.html#upgrade0.90
Also if you want to stay on CDH, CDH3b4 was just released.
J-D
On Thu, Feb 24, 2011 at 4:08 PM, Dmitriy Lyubimov dlie...@gmail.com wrote:
Thank you, sir.
Do you happen to know if hbase swtich from 0.89 CDH3b3 to 0.90.1 is going to
be
It could be due to slow splits, heavy GC, etc. Make sure your machines
don't swap at all, that HBase has plenty of memory, that you're not
trying to use more CPUs than your machines actually have (like setting
4 maps on a 4 cores machine when also using hbase), etc.
Also upgrading to 0.90.1 will
if it fixes it.
J-D
On Wed, Feb 23, 2011 at 11:18 AM, charan kumar charan.ku...@gmail.com wrote:
Hi J-D,
There are no NPE's in the log.
Thanks,
Charan
On Wed, Feb 23, 2011 at 11:04 AM, Jean-Daniel Cryans
jdcry...@apache.orgwrote:
Check you master log, if you see a lot of NPEs then it means
the .oldlogs folder without
causing any issues?
I will also look into upgrading..
On Wed, Feb 23, 2011 at 11:23 AM, Jean-Daniel Cryans
jdcry...@apache.orgwrote:
I'll have to trust you on that :)
The other possible situation is that you are inserting a ton of data
and logs are generated
How do you create the configuration object Dan? Are you doing:
Configuration conf = HBaseConfiguration.create();
Job job = new Job(conf, somename);
or are you just creating a normal Configuration?
BTW the code I wrote is what I expect people do and what I'm doing myself.
J-D
On Wed, Feb 23,
to include the HBase
configuration too then it would work fine.
Thanks,
On 23 Feb 2011, at 19:44, Jean-Daniel Cryans jdcry...@apache.org wrote:
How do you create the configuration object Dan? Are you doing:
Configuration conf = HBaseConfiguration.create();
Job job = new Job(conf, somename
(moving to the user ML)
in the shell (bin/hbase shell), type: help create
And then you will have the help for creating table along with examples.
Regarding the actual modeling of your table, putting the userid as the
row key (no need for a family) could be a viable option but the rest
seems
:a/1298037767127/Put/vlen=3,
row3/family1:b/1298037770111/Put/vlen=3,
row3/family1:c/1298037774954/Put/vlen=3}
I see there is everything but value. What should I do to get value
on stdin too?
Ondrej
On 02/18/11 20:01, Jean-Daniel Cryans wrote:
You have a typo, it's
Ted asked about the JVM version but I don't think you answered that.
In any case, try with u17.
J-D
On Sat, Feb 19, 2011 at 3:58 AM, Enis Soztutar enis.soz.nu...@gmail.com wrote:
Yes indeed but no luck.
Enis
On Fri, Feb 18, 2011 at 11:50 AM, Jean-Daniel Cryans jdcry...@apache.org
wrote
This exception happens when hbase.zookeeper.property.clientPort cannot
be found from any file in the classpath (the bit about zoo.cfg is a
bit confusing I agree).
If you didn't change it, then it should be found in the
hbase-default.xml file which is now packaged inside the hbase jar
since 0.90.0
For Hadoop, from the website's documentation:
http://hbase.apache.org/notsoquick.html#hadoop
and regarding zookeeper use the one that's ship with HBase.
J-D
On Tue, Feb 22, 2011 at 2:53 PM, Oleg Ruchovets oruchov...@gmail.com wrote:
Hi ,
which hadoop and zookeeper version should I use with
On Thu, Feb 17, 2011 at 1:50 AM, Jean-Daniel Cryans jdcry...@apache.org
wrote:
Actually I never thought of having a special case for that... and I
don't see any jira about it. Would you mind opening a new one for
that, I think it's a good idea for those times when you're developing
something
A Get is a random read, so expect it to be slower than let's say a
scanner or a random insert (the other calls that are made in your
code). Unless you are able to keep all that data in the block cache of
the region servers, those calls are going to be expensive.
A change that would be very easy
, Feb 18, 2011 at 3:58 PM, Jean-Daniel Cryans jdcry...@apache.org
wrote:
There's probably (and I'm 99% sure) a DNS timeout happening when
resolving your machine's hostname. Review your DNS settings.
J-D
On Fri, Feb 18, 2011 at 10:53 AM, Fabiano D. Beppler f...@knewin.com
wrote:
Hi,
I am
a lot. Is this optimized for a specific JVM or does
it not matter?
On Fri, Feb 18, 2011 at 5:46 PM, Todd Lipcon t...@cloudera.com wrote:
On Fri, Feb 18, 2011 at 12:10 PM, Jean-Daniel Cryans
jdcry...@apache.orgwrote:
The bigger the heap the longer the GC pause of the world when
There's probably (and I'm 99% sure) a DNS timeout happening when
resolving your machine's hostname. Review your DNS settings.
J-D
On Fri, Feb 18, 2011 at 10:53 AM, Fabiano D. Beppler f...@knewin.com wrote:
Hi,
I am running a very simple JUnit test with HBase and the test takes a lot of
time
Less regions, but it's often a good thing if you have a lot of data :)
It's probably a good thing to bump the HDFS block size to 128 or 256MB
since you know you're going to have huge-ish files.
But anyway regarding penalties, I can't think of one that clearly
comes out (unless you use a very
Just to make sure, you did check in the .out file after a failure right?
J-D
On Thu, Feb 17, 2011 at 10:14 PM, Enis Soztutar
enis.soz.nu...@gmail.com wrote:
Hi,
Thanks everyone for the answers.
I had already increase the file descriptors to 32768. The region servers
and the zookeeper
The master should finish processing those dead servers at some point
and it seems it's not happening? Unfortunately without the log nobody
can'tell why. If you can post the complete log in pastebin or put it
on a web server then we could take a look.
J-D
On Fri, Feb 18, 2011 at 12:39 AM, Yi
of RAM (48GB) so that should not be
an issue - I just want to minimize the risk that GC will cause problems.
thanks again.
-chris
On Feb 18, 2011, at 11:59 AM, Jean-Daniel Cryans wrote:
That's what I usually recommend, the bigger the flushed files the
better. On the other hand, you only
findings...
Vishal
On Wed, Feb 16, 2011 at 8:00 PM, Jean-Daniel Cryans jdcry...@apache.org
wrote:
It's best to have different families for data of different nature and
when you usually don't read/write them together. For sure it shouldn't
slow you down as much as it does (because of HBASE-3149
First, loading into 3 families is currently a bad idea and is bound to
be inefficient, here's the reason why:
https://issues.apache.org/jira/browse/HBASE-3149
Those log lines mean that your scanning of the first table is
generating a log of block cache churn. When setting up the Map, set
your
Hi Otis,
Excellent reflexion, unfortunately I don't think anyone benchmarked it
to give a definitive answer.
One thing I'm sure of is that worse than screwing up the OS cache, it
also screws up the block cache! But this is the price to pay to clear
up old versions and regroup all store files
cache's being screwed? Or because it
also requires a full CPU to major compact?
The answer is probably all of the above.
J-D
On Wed, Feb 16, 2011 at 10:03 AM, Jean-Daniel Cryans
jdcry...@apache.org wrote:
Hi Otis,
Excellent reflexion, unfortunately I don't think anyone benchmarked it
to give
families so that I can have same qualifiers in them. and also
the data in those families are LIVE:MasterA:MasterB
Vishal
On Wed, Feb 16, 2011 at 6:22 PM, Jean-Daniel Cryans
jdcry...@apache.orgwrote:
Very often there's no need for more than 1 family, I would suggest you
explore
in 0.20.6 is it?
On Wed, Feb 16, 2011 at 2:30 PM, Jean-Daniel Cryans jdcry...@apache.org
wrote:
To disable a region, it's memstore must first be flushed. That's done
sequentially for all regions that must be disabled inside a table.
There's not really a way around it unless you don't need
here is speed for inserts and I am going
to definitely give a try to single family approach.
Vishal
On Wed, Feb 16, 2011 at 6:53 PM, Jean-Daniel Cryans
jdcry...@apache.orgwrote:
I don't understand... is having the same qualifier a hard requirement?
Worst case you could have a prefix.
J-D
Start with this:
http://www.cloudera.com/blog/2010/08/hadoophbase-capacity-planning/
Then regarding the number of servers... it's really hard to tell,
you'd have to test with a handful of machines first and see how they
perform under your type of load. Scaling is then as easy as adding the
new
Compactions are done in the background, they won't block writes.
Regarding splitting time, it could be that it had to retry a bunch of
times in such a way that the write timed out, but I can't say for sure
without the logs.
Have you considered using the bulk loader? I personally would never
try
On Tue, Feb 15, 2011 at 11:24 AM, Chris Tarnas c...@email.com wrote:
We are definitely considering writing a bulk loader, but as it is this fits
into an existing processing pipeline that is not Java and does not fit into
the importtsv tool (we use column names as data as well) we have not
withstand a 120 second
pause?
thanks,
-chris
On Feb 15, 2011, at 1:37 PM, Chris Tarnas wrote:
On Feb 15, 2011, at 11:32 AM, Jean-Daniel Cryans wrote:
On Tue, Feb 15, 2011 at 11:24 AM, Chris Tarnas c...@email.com wrote:
We are definitely considering writing a bulk loader
.
On Feb 15, 2011, at 2:10 PM, Jean-Daniel Cryans wrote:
Yeah if it's the same key space that splits, it could explain the
issue... 65 seconds is a long time! Is there any swapping going on?
CPU or IO starvation?
In that context I don't see any problem setting the pausing time higher.
J-D
speed would only be acceptable if you batch up -- I understand what you
are talking about here (without batching-up, HBase simply become very
sluggish). Can you comment if Cassandra needs a batch-up mode? (I recall
Twitter said they just keep putting results into Cassandra for its analytics
Please use the hbase mailing list for HBase-related questions.
Regarding your issue, we'll need more information to help you out.
Haven you checked the logs? If you see exceptions in there, did you
google them trying to figure out what's going on?
Finally, does your setup meet all the
This is all in the documentation:
http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/mapreduce/package-summary.html#classpath
J-D
On Mon, Feb 14, 2011 at 3:00 AM, Hari Sreekumar
hsreeku...@clickable.com wrote:
I have this entry in hadoop-env.sh
export
Erm well could it be that your script is truncating a bunch of tables
whereas in the shell it only does one? Apart from that I can't think
of anything obvious, the truncate command really does the same thing
as you.
J-D
On Mon, Feb 14, 2011 at 9:43 AM, Peter Haidinyak phaidin...@local.com wrote:
Hey Bradford,
Could we see the full thing? I bet there's a bunch of ERROR. Look for
the dumping of metrics (grep for dump) and then get all the lines
before that (until you see it's doing normal stuff).
BTW that log is telling me that another region server died before that one.
J-D
On Sun, Feb
There are scores of issues that were fixed in 0.89 and 0.90 around
that part of the code, so it's really hard to tell if you're hitting
any of those. My recommendation is to upgrade... none of the big
installations that I know of are on 0.20.6
J-D
On Sat, Feb 12, 2011 at 7:02 AM, Murali Krishna.
Same answer that I gave to your other email:
We'll need more information to help you out.
Have you checked the logs? If you see exceptions in there, did you
google them trying to figure out what's going on?
Finally, does your setup meet all the requirements?
hbase list 'abc.*'
hbase(main):002:0
--
2011/2/14 Jean-Daniel Cryans jdcry...@apache.org
Same answer that I gave
There's the split command in the shel.
HBaseAdmin has that same method.
In the table's page from the master's web UI, there's a split button.
Finally, when creating a table, you can pre-specify all the split keys
with this method:
I think not, it opens and edits .META. so it would be like having
multiple region servers serving it (which is always bad).
J-D
On Thu, Feb 10, 2011 at 5:22 AM, Sebastian Bauer ad...@ugame.net.pl wrote:
Hi, is anybody knows that ./bin/hbase org.apache.hadoop.hbase.util.Merge
can run in
The first thing to do would be to look at the datanode logs a the time
of the outage. Very often it's caused by either ulimit or xcievers
that weren't properly configured, checkout
http://hbase.apache.org/notsoquick.html#ulimit
J-D
On Thu, Feb 10, 2011 at 10:42 AM, Venkatesh
works on an offline cluster, it goes straight to
the META HFiles, thus cannot be run in parallel.
It shouldn't be too hard to hack up Merge to work on an online
cluster, offline table.
On Thu, Feb 10, 2011 at 10:09 AM, Jean-Daniel Cryans
jdcry...@apache.org wrote:
I think not, it opens
You should take a look at the master log and see if it looks normal or
not. Maybe also check if the process is running.
J-D
On Sat, Feb 5, 2011 at 11:43 AM, Jérôme Verstrynge jvers...@gmail.com wrote:
Hi,
I have installed Cloudera's CDH3 successfully on a node. I have written a
small
Your HDFS doesn't look healthy to me, it doesn't accept new blocks!
Check the datanode / namenode logs and see if any exception in there.
J-D
On Mon, Feb 7, 2011 at 12:58 PM, Karthik Manimaran
karthikmanima...@gmail.com wrote:
Hi,
My HBase setup was running fine for a couple of months and all
UUIDs could be used for that in general.
What's your row key currently made of? Kind of hard to help you if we
don't know anything about your use case :)
J-D
On Sun, Feb 6, 2011 at 10:15 PM, som_shekhar kolluru.sha...@wipro.com wrote:
Hi All,
I would like to know how to provide the unique
:59)
at org.apache.hadoop.hbase.master.LogCleaner.chore(LogCleaner.java:140)
at org.apache.hadoop.hbase.Chore.run(Chore.java:66)
at org.apache.hadoop.hbase.master.LogCleaner.run(LogCleaner.java:167)
On Sat, Jan 29, 2011 at 8:43 PM, Jean-Daniel Cryans
jdcry
On the scan, you can setCaching with the number of rows you want to
pre-fetch per RPC. Setting it to 2 is already 2x better than the
default.
J-D
On Thu, Feb 3, 2011 at 1:35 PM, Something Something
mailinglist...@gmail.com wrote:
After adding the following line:
Inline.
J-D
1.
I posted a question a couple days ago about raid configuration for Hadoop and
the answer is JBOD however, once you setup that up and you are going through
your linux install what volume formatting do you select? ext3/4 lvm?
ext4 seems to be the new favorite, before that
I don't think HBase is really needed here, unless you somehow need
random read/write to those search queries.
J-D
On Wed, Feb 2, 2011 at 1:27 PM, Peter Haidinyak phaidin...@local.com wrote:
Hi all,
I was just tasked to take the keywords used for a search and put them
in HBase so we
(moving this to the user mailing list where it belongs)
You need to make sure that your webapp knows the address of the
JobTracker, usually this is done by either putting mapred-site.xml on
your app's classpath or you can set mapred.job.tracker correctly so
that in createSubmittableJob you would
(moving to the user mailing list, where it belongs)
My educated guess is that you had a GC pause that lasted for more than a
minute while a file was being written to. Even if the write wasn't
happening, your region server would have committed suicide anyways since it
was probably past it's lease
There's some sort of rate limiting for file deletion, I think it's 20 every
time it runs (which is every minute). Could it be that your region servers
are creating them faster than that?
In any case, it's safe to delete them but not the folder itself. Also please
open a jira and assign it to me.
Cool, so the coprocessor will feed the value in the database to me and the
value that is coming in just before it is written?
With bytes, I am using serialized json so the example still applies perfectly
where I could merge the results in the coprocessor and the coprocessor writes
the
Use the normal Java API in your job to do that e.g. HTable.
J-D
On Thu, Jan 27, 2011 at 7:48 AM, Hiller, Dean (Contractor)
dean.hil...@broadridge.com wrote:
I have a map/reduce job I coded up to play with this and there is this
line that sets the table
hprof files are heap dumps from java when you get an OutOfMemoryError,
it's generated because HBase ships with
-XX:+HeapDumpOnOutOfMemoryError in conf/hbase-env.sh
J-D
On Thu, Jan 27, 2011 at 9:25 AM, Peter Haidinyak phaidin...@local.com wrote:
Hi,
My cluster went do because some HPROF files
For HBase it's all byte[], so there's currently nothing as involved as
what you describe. What happens is that the second Put will overshadow
the first one, but both are accessible.
In the future I guess you could implement a coprocessor that does that
kind of check, but it's only available in
To mitigate heap fragmentation, you could consider adding more nodes
to the cluster :)
Regarding rolling restarts, currently there's one major issue:
https://issues.apache.org/jira/browse/HBASE-3441
How it currently works is a bit dumb, when you cleanly close a region
server it will first close
is
gone.
J-D
On Thu, Jan 27, 2011 at 4:13 PM, Ted Yu yuzhih...@gmail.com wrote:
Should steps 1 and 2 below be exchanged ?
Regards
On Thu, Jan 27, 2011 at 3:53 PM, Jean-Daniel Cryans
jdcry...@apache.orgwrote:
To mitigate heap fragmentation, you could consider adding more nodes
to the cluster
Not as far as I know, unless you disabled splits from the beginning
like some ppl do.
J-D
On Thu, Jan 27, 2011 at 4:22 PM, Ted Yu yuzhih...@gmail.com wrote:
Is there a way to disable splitting (on a particular region server) ?
On Thu, Jan 27, 2011 at 4:20 PM, Jean-Daniel Cryans
jdcry
that the maximum number of regions for the soon-to-go region
server is 0?
On Thu, Jan 27, 2011 at 3:53 PM, Jean-Daniel Cryans
jdcry...@apache.orgwrote:
Writing this gives me an idea... I think one easy way we could
achieve this region draining problem is by writing a jruby script
that:
1
I happens on this call:
https://github.com/apache/hadoop/blob/trunk/src/core/org/apache/hadoop/conf/Configuration.java#L1002
J-D
On Wed, Jan 26, 2011 at 10:43 AM, Weishung Chung weish...@gmail.com wrote:
It seems like the properties are only lazily initialized from xmls on the
first call to
It seems to be coming from the region server side... so one thing you
can check is the region server logs and see if the NPEs are there. If
not, and there's nothing suspicious, then consider enabling DEBUG for
hbase and re-run the job to hopefully get more information.
J-D
On Wed, Jan 26, 2011
This feature doesn't exist in HBase at the moment.
J-D
On Wed, Jan 26, 2011 at 7:22 AM, Weishung Chung weish...@gmail.com wrote:
I want to group tables in different databases, could I create database and
have tables under it as in mysql?
Thank you
Caching is the number of rows that will be fetched per RPC, depending
on how big your rows are you might want to set it larger or smaller.
Try 10 then do some experiments.
There's not that many more, speed of reading is always improved with
caching. Make sure your data can fit in the block cache
Inline.
J-D
On Tue, Jan 25, 2011 at 3:48 AM, 陈加俊 cjjvict...@gmail.com wrote:
One programme scan from regions[0].startKey and stop at regions[0].endKey
,and another programme scan from scan from regions[1].startKey and stop at
regions[1].endKey.
Every programme get the row by scan and then
The master splits the logs per region before reassigning them. The log
splits are put directly in the region's folder so that when a region
server opens a region that comes from a dead server, it looks for
those files and processes them first before opening the region (to
ensure consistency).
If the region servers gets YouAreDeadException, it does an abort and
won't flush the data since another region server could already be
serving the region. If you're not writing to the WAL, then yes it's
data loss.
Not sure what you mean by shuts down cleanly in your case, if you
see a log that
Although it should handle that NPE, the fact that
ReplicationLogCleaner is running without replication being enabled
means that your hbase-default.xml is from an older version. Please
refresh it.
J-D
2011/1/20 Yifeng Jiang yifeng.ji...@mail.rakuten.co.jp:
Hi,
I'm using HBase-0.90 and I found
The Placement in byte shows some value like this: [B@298488ef
This line doesn't prove that there's actual data in that byte array,
it only prints the object's address. You can verify there's data by
looking at the length.
Considering that the value object is empty, it'd suggest that you
HBase doesn't support namespacing at the moment, but what I've seen in
the past is having a configuration file for your application that has
the configurations for each environment and some class that does all
that managing by using a string to specify to which cluster you which
to talk to.
echo 'some command' | bin/hbase shell
J-D
On Wed, Jan 19, 2011 at 2:01 PM, Nanheng Wu nanhen...@gmail.com wrote:
Hi,
Sorry for the stupid question. I want to execute some hbase shell
commands like list or create table from the command line directly,
instead of through the interactive hbase
You don't have to, but it's best to do it. This will help you
understanding why:
http://www.larsgeorge.com/2010/05/hbase-file-locality-in-hdfs.html
J-D
On Wed, Jan 12, 2011 at 10:04 AM, Peter Haidinyak phaidin...@local.com wrote:
Hi,
This might be a really dumb question but do you need to run
] On Behalf Of Jean-Daniel
Cryans
Sent: Wednesday, January 12, 2011 10:12 AM
To: user@hbase.apache.org
Subject: Re: Region Server on Data Node
You don't have to, but it's best to do it. This will help you
understanding why:
http://www.larsgeorge.com/2010/05/hbase-file-locality-in-hdfs.html
J
, 2011 at 10:22 AM, Jean-Daniel Cryans
jdcry...@apache.orgwrote:
The region server knows nothing about the file locality. The magic
happens between the DFSClient and the Namenode; in HDFS, new files
will have one block on the local datanode when it's possible, but
existing ones won't be moved
to provide a patch, after
agreement on where to put this option. Should it be member of Mutation, or a
new call should be created (eg. mutateRowsNoWal())?
Thanks,
Jan
On 7.1.2011 19:21, Jean-Daniel Cryans wrote:
Not by design, it's really just missing.
J-D
On Fri, Jan 7, 2011 at 4:56
HBaseHUT is used to solve he Get+Put problem, so if it's your problem
as well then do look into it.
To answer your first question, that method will group Puts by region
server meaning that it will do anywhere between 1-n where n is the
number of RS, and that's done in parallel.
J-D
On Mon, Jan
Inline.
*Region/Meta Cache
*
Often times the region list is not hot and thrift has to talk to the meta
table. We have 6k+ regions and growing quickly and expect 1k+/node. Can we
help our performance by pre-caching all region locations? How many regions
can thrift keep before over-writing in
, would you recommend the use of MapReduce to accomplish the samething?
On Mon, Jan 10, 2011 at 11:38 AM, Jean-Daniel Cryans
jdcry...@apache.orgwrote:
HBaseHUT is used to solve he Get+Put problem, so if it's your problem
as well then do look into it.
To answer your first question
lotsa rows? That's 1k or 1B? Inside a OLTP system or OLAP?
J-D
On Mon, Jan 10, 2011 at 10:58 AM, Weishung Chung weish...@gmail.com wrote:
Jonathan, awesome, best of breed APIs!
Jean, I would like to insert lotsa new rows with many columns in a
particular column family* **programmatically in
Is this the same as the hregion.max.filesize setting?
Yes.
If HConnectionManager is doing it is it then hbase memory of Thrift server
memory that caches the region locations?
That's the same JVM, so same memory? I'm not sure what you're asking
about but if it can help, you give more memory
simultaneously. I want these multiple batches of
insertions to be done super fast. At the same time, I would like to be able
to scale up to 100k rows at a time (the goal). Now, I am building a cluster
of size 6 to 7 nodes.
On Mon, Jan 10, 2011 at 1:03 PM, Jean-Daniel Cryans
jdcry
-Original Message-
From: jdcry...@gmail.com [mailto:jdcry...@gmail.com] On Behalf Of Jean-Daniel
Cryans
Sent: Monday, January 03, 2011 4:31 PM
To: user@hbase.apache.org
Subject: Re: batch reads of columns?
I would be tempted to get a taller table table instead of a very very
wide one, scanning
, or is there any criteria? because when I
decrease number of tasks.maximum to run same job, it's ok.
Shen
On Fri, Jan 7, 2011 at 3:20 AM, Jean-Daniel Cryans jdcry...@apache.orgwrote:
Shen,
It's a design decision, and we historically preferred to let cluster
managers decide whether they want to restart
Not by design, it's really just missing.
J-D
On Fri, Jan 7, 2011 at 4:56 AM, Jan Lukavský
jan.lukav...@firma.seznam.cz wrote:
Hello everyone,
we are missing possibility to disable WAL through Thrft server, is this
option missing by design?
Thanks,
Jan
Shen,
It's a design decision, and we historically preferred to let cluster
managers decide whether they want to restart the processes that died
or investigate why it has died then decide on what they want to do.
You can easily write tools that will restart the region servers if
they die, but the
With one cluster you really only need one, and it doesn't seem to be
running from what I can tell:
2011-01-05 15:20:12,185 WARN zookeeper.ClientCnxn
(ClientCnxn.java:run(967)) - Exception closing session 0x0 to
sun.nio.ch.selectionkeyi...@561279c8
java.net.ConnectException: Connection refused
Maybe it's not clear, by need to be in the same timezone it means
that the machines' configured timezone needs to be the same. So yes
they can be in different physical timezones but it will work if the
configured tz is UTC.
J-D
On Wed, Jan 5, 2011 at 11:10 AM, Joe Pallas pal...@cs.stanford.edu
Oh then I guess I was mistaken during all that time (makes a lot of
sense tho now that I think about it).
J-D
On Wed, Jan 5, 2011 at 11:41 AM, Joe Pallas pal...@cs.stanford.edu wrote:
On Jan 5, 2011, at 11:17 AM, Jean-Daniel Cryans wrote:
Maybe it's not clear, by need to be in the same
I don't have a whole lot of recent HBase on EBS experience, but when I
did do it my main issue was that sometimes some EBS volumes would
become unavailable.
The way I see it is that you have an additional moving part in your
whole stack, thus there's a chance it will generate a new set of
First, never swap with Java. Disable it on your machines.
Second, to go at the bottom of this issue you need to go where it
starts showing exceptions in the logs. In your case it seems we only
see indirect symptoms of a forceful failover by the HBase master.
Somewhere before that there should be
901 - 1000 of 1248 matches
Mail list logo
