3. if pool is out:
I get a normal Htable instance, because HTablePool.getTable will call
findOrCreateTable(), once there is not table available, it creates a
HbaseTable instance instead of PooledHbase instance in this case, once I am
done with query, I call Htable.close(), it will just release
Here it is:
https://issues.apache.org/jira/browse/HBASE-6856
On 09/21/2012 08:42 PM, Harsh J wrote:
Daniel,
Nice follow up! We could add some notes around these to the doc as
well. Please do post back a JIRA link once you've filed it.
On Thu, Sep 20, 2012 at 8:15 PM, Daniel Iancu
h of the error
cases - UnknownScannerException and LeaseException. Mind filing a
JIRA? :)
On Thu, Sep 20, 2012 at 7:21 PM, Daniel Iancu wrote:
Thaaank you! I was waiting for this email for months. I've read all the
posts regarding lease timeouts and see that people usually have them for 2
reasons. One,
Thaaank you! I was waiting for this email for months. I've read all the
posts regarding lease timeouts and see that people usually have them for
2 reasons. One, the normal case where the client app does not process
the row fast enough so they get UnknownScannerException and some had the
issue b
Hi Oleg
Normally, if your client does not process rows fast enough you should
get a UnknownScannerException. A higher cache value means a higher time
between two next() operations (that renew the lease) so a higher chance
to get a USE. Try to optimize the row processing in the client and lower
Sorry for the bad typing, Don't use HTablePool.putTable()...
The class is not abandoned from what I know, only the putTable method
should be avoided.
On 03/14/2012 10:32 AM, Daniel Iancu wrote:
Hi
Don't use HTablePut.put() method, just HTable.close on the table you got
from pool w
Hi
Don't use HTablePut.put() method, just HTable.close on the table you got
from pool when you no longer need it. Don't worry, the table will be put
back in the pool after you closed it.
On 03/14/2012 10:14 AM, 韶隆吴 wrote:
Hi all:
From blogs I know HTablePool class is like ConnectionPool.B
e the hadoop M/R job ?
Thanks
Haijia
--
Daniel Iancu
Java Developer,Big Data Solutions Romania
1&1 Internet Development srl.
18 Mircea Eliade St
Sect 1, Bucharest
RO Bucharest, 012015
www.1and1.ro
Phone:+40-031-223-9081
I deployed it pretty easy on our internal repo by checking out the tag
0.92.0 (I assume this is the release) and *mvn deploy -DskipTests=true*.
Or you can move tests to a separate module eg hbase-test and add a
dependency to hbase. If all tests in hbase-test pass then you can
release the hbase
Hi Alex
Do you use a filter for KVs in scanner? How many columns per row can you
have?
Thanks
Daniel
On 02/12/2012 09:55 PM, Alex Baranau wrote:
forgot to mention that caching on Scanner was set to only 64 and I the
estimate for records size is< 100KB
Hi
Looks like Hbase does not support the skip of bad records from MR. I've seen
that HBASE-1239 was invalidated and there is not much talk around it.
In hour case that would have been a useful feature right now, are there any
alternatives to this or do we need to build ourselves some mechanism t
If you move the region to another host, do you same same perf (Perhaps
some hardware issue?).
Done more testing today. Is not related to a particular region, it
happened today with an other region on same machine. Also is not a
permanent issue, after some time I retried and the scan was fast a
Hi there
I'm investigating a problem we have with a MR job and I discovered that
the tasks that fail (scan lease expired while fetching next row) were
processing one particular region.
I've written a small app that scans that region and counts its rows and
run it on same machine where region is
Will you add this to Hbase manual ? Does it worth trying 24, 32 Gb per
RS under heavy load?
On 12/07/2011 03:20 AM, Doug Meil wrote:
Folks at recent Hbase
hackathons were talking about not going higher than 16Gb for RS.
logging the exception produced by the hbase.rpc.timeout on
the client side on WARN, not debug like it is right now.
That makes sens. Mind making an issue (and adding a patch if inclined)?
The above seems more painful than it need be figuring what was going on.
Thanks boss,
St.Ack
--
Daniel
I've edited the JIRA description to be more formal
On 07/01/2011 07:10 PM, Ted Yu wrote:
When logging JIRA, please use formal writing instead of pasting contents of
email directly.
I will post a comment on the JIRA related to PooledHTable.close().
On Fri, Jul 1, 2011 at 9:06 AM, Daniel
, wouldn't it be good idea to add close() to
HTablePoolEnhanced so that underlying HTables can be closed by client apps as
part of application shutdown.
Regards,
Srikanth
-Original Message-
From: Daniel Iancu [mailto:daniel.ia...@1and1.ro]
Sent: Friday, July 01, 2011 7:55 PM
To:
actually return the table to pool transparently.
Thanks
On Fri, Jul 1, 2011 at 7:25 AM, Daniel Iancu wrote:
Hi
It look like the usage of HTablePool might be improved. Now, once you get
the connection from pool you must take good care to return it by calling
HTablePool.putTable(table);
If you
Hi
It look like the usage of HTablePool might be improved. Now, once you
get the connection from pool you must take good care to return it by
calling HTablePool.putTable(table);
If you close the table (say, you don't read carefully the Javadoc)
your htable will not be reused.
Other case migh
Hello everybody
I've run into this strange problem. We run a 6 RS cluster and suddenly
the client application started reporting errors, region not online. In
the web console all regionserver appeared up. I've run hbck and got
strange results
Number of Tables: 2
Number of live region servers
e for now is to extend TableSplit and to use the
subclass.
I use cloudera hbase cdh3u0 version.
Do you know about this problem? Should I open a Jira issue for that, or it
already exists?
Thanks,
Lucian
--
Daniel Iancu
Java Developer,Web Components Romania
1&1 Internet Development s
version of this patch for CDH3b3/4.
St.Ack
On Thu, Apr 7, 2011 at 9:35 AM, Daniel Iancu wrote:
Hello everybody
We've run into this, now popular, error on our cluster
2011-04-07 16:28:00,654 WARN IPC Server handler 0 on 8020
org.apache.hadoop.hdfs.StateChange - DIR* NameSystem.startFile: f
file
# start master, and watch it replay the log.
This appears to break the lease as new .log file does not have this issue.
-Jack
On Thu, Apr 7, 2011 at 9:35 AM, Daniel Iancu wrote:
Hello everybody
We've run into this, now popular, error on our cluster
2011-04-07 16:28:00,654 WARN IPC S
Hello everybody
We've run into this, now popular, error on our cluster
2011-04-07 16:28:00,654 WARN IPC Server handler 0 on 8020
org.apache.hadoop.hdfs.StateChange - DIR* NameSystem.startFile: failed
to create file
/hbase/.logs/search-hadoop-eu001.v300.gmx.net,60020,1302075782687/search-hadoop
ction to 'desktop,60020'. Later
we refer to an IP for desktop. Is your dns set so resolve and reverse
lookup give same answer? HBase 0.90.x is finicky in this regard (To
be fixed).
St.Ack
On Fri, Mar 4, 2011 at 6:39 AM, Daniel Iancu wrote:
Hi
I've updated our dev environment fr
Hi
I've updated our dev environment from Hbase 0.90.0 (ASF+CDH3b3) which
behaved very stable to Hbase 0.90.1 (CDH3B4) and since then the HMaster
dies regularly. Issue seems to be regarded to the connection to
Zookeeper. Even if I use a standby HMaster, this one also dies from same
cause:
20
Hi Stack
I've looked again over our conversation and I think my comment was not
very clear. So, your pom does
not need to be patched, it is fine as long as people don't use
hbase-test library in their projects.
If they do, again pom is fine but their pom need a dependency on the
right hadoop-te
Sure, I can do that. I was thinking, maybe somehow the HBase Mini Cluster class
can be refactored so it hides the dependency on the Hdfs Mini Cluster class.
Something like instead of us instantiating the HDFS mini cluster and then
injecting it in the HBase minicluster we can tell the HBase clust
05 PM, Daniel Iancu wrote:
Hi Stack
Since project is buildable you'll need to configure the distribution
management stuff. You could use an empty project to try and error the
settings then use them for HBase. If it helps, I can try deploying it on our
internal repository and then share the step
annot disable tests for release, just for install.
Regards
Daniel
On 01/26/2011 05:35 PM, Stack wrote:
On Wed, Jan 26, 2011 at 5:00 AM, Daniel Iancu wrote:
Hi there
Now that 0.90 is out and 'HBASE-1933 Upload Hbase jars to a public maven
repository' is in changes file is there a repo wh
/projects/hbase-0.89.20100924/src/main/java/org/apache/hadoop/hbase/master/RegionManager.java:[690,9]
cannot find symbol
symbol : class NotAllMetaRegionsOnlineException
location: class org.apache.hadoop.hbase.master.RegionManager
And tons more.
--
Daniel Iancu
Java Developer,Web Components Roman
find
correct
answer.
Thank a lot for support.
--
Daniel Iancu
Java Developer,Web Components Romania
1&1 Internet Development srl.
18 Mircea Eliade St
Sect 1, Bucharest
RO Bucharest, 012015
www.1and1.ro
Phone:+40-031-223-9081
Email:daniel.ia...@1and1.ro
IM:dia...@united.domain
prohibited. If you receive this e-mail in error, please
notify the sender by
phone or email immediately and delete it!
--
Daniel Iancu
Java Developer,Web Components Romania
1&1 Internet Development srl.
18 Mircea Eliade St
Sect 1, Bucharest
RO Bucharest, 012015
www.1and1.ro
Phone:+4
33 matches
Mail list logo
