>A flush every time you add to BufferedMutator?
Yes
> That will result in an rpc
I realize this fact.
>let the BufferedMutator backing buffer fill and auto flush
It's my next step :) Just wanted to to get rid of exception and see how it
goes.
>may want to consider using Table.put(List puts) ins
If you have discrete sets of Puts that happen across different threads, you
may want to consider using Table.put(List puts) instead of
BufferedMutator.
-Solomon
On Wed, Feb 10, 2016 at 6:58 PM, Stack wrote:
> A flush every time you add to BufferedMutator? That will result in an rpc
> each time.
A flush every time you add to BufferedMutator? That will result in an rpc
each time. Is that what you want? You don't want to let the BufferedMutator
backing buffer fill and auto flush?
Be sure to call close when shutting down else whatever is in the
BufferedMutator backing buffer will be lost.
S
It helped, thanks.
Now I have single reusable BufferedMutator instance and I don't call
.close() after mutations, I call .flush() Is it ok?
2016-02-09 23:09 GMT+01:00 Stack :
> On Tue, Feb 9, 2016 at 9:46 AM, Serega Sheypak
> wrote:
>
> > I've modified my code:
> >
> > void saveUsers(Collection
On Tue, Feb 9, 2016 at 9:46 AM, Serega Sheypak
wrote:
> I've modified my code:
>
> void saveUsers(Collection users){
> if(users && !users.isEmpty()){
>
> // Get Connection instance, instance created once. , BM is new
> for each request.'
>
If new for each request, don't both
Sorry, occasionally sent.
Now exception says about *sendMultiAction*, before that is wan't multi.
Feb 09, 2016 8:45:17 PM *org.apache.hadoop.hbase.client.AsyncProcess
sendMultiAction*
WARNING: #127, the task was rejected by the pool. This is unexpected.
Server is node02.admanager.pro,60020,1447
I've modified my code:
void saveUsers(Collection users){
if(users && !users.isEmpty()){
// Get Connection instance, instance created once. , BM is new
for each request.
def mutator =
getConnection().getBufferedMutator(getBufferedMutatorParams())
List
On Tue, Feb 9, 2016 at 6:37 AM, Serega Sheypak
wrote:
> Hi, thanks for reply!
>
>
> > What should we add here to make the doc more clear on BufferedMutator?
> >
>
> http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/client/BufferedMutator.html
> It's pretty clear.
>
> > And in the example?
>
Hi, thanks for reply!
> What should we add here to make the doc more clear on BufferedMutator?
>
http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/client/BufferedMutator.html
It's pretty clear.
> And in the example?
Example shows that both connection and buffered mutator are thread safe an
On Mon, Feb 8, 2016 at 3:01 PM, Serega Sheypak
wrote:
> Hi, I'm confused with new HBase 1.0 API. API says that application should
> manage connections (Previously HConnections) on their own. Nothing is
> managed itnernally now.
>
>
That is right.
> Here is an example:
>
> https://hbase.apache
bq. users.each{ mutator.mutate(toPut(it))}
I assume users is collection of User's.
Have you tried obtaining / closing mutator for each User instead of sharing
the mutator ?
See TestClientNoCluster for sample usage.
On Mon, Feb 8, 2016 at 3:01 PM, Serega Sheypak
wrote:
> Hi, I'm confused with n
Hi, I'm confused with new HBase 1.0 API. API says that application should
manage connections (Previously HConnections) on their own. Nothing is
managed itnernally now.
Here is an example:
https://hbase.apache.org/xref/org/apache/hadoop/hbase/client/example/BufferedMutatorExample.html
It gives no
12 matches
Mail list logo
