Seems the problem only exists when using load balance sink processor with
groups of sinks that use lzo compression.
Looking into the details.
2015-09-08 10:19 GMT+08:00 Shady Xu :
> Using different prefixes does not fix the problem. Any other idea?
>
> 2015-09-07 10:19 GMT+08:00 Shady Xu :
>
>>
I got Flume configured to read Avro events from Kafka source and I'm also
attaching the schema like this:
~~~
f1.sources.kafka-source.interceptors.attach-f1-schema.type = static
f1.sources.kafka-source.interceptors.attach-f1-schema.key =
flume.avro.schema.url
f1.sources.kafka-source.interceptors.a
Thanks for your answer.
Aljoscha
On Tue, 8 Sep 2015 at 20:04 Johny Rufus wrote:
> Your assumption is correct, as duplicates in a failure scenario will occur.
>
> Thanks,
> Rufus
>
> On Tue, Sep 8, 2015 at 4:10 AM, Aljoscha Krettek
> wrote:
>
>> Hi,
>> as I understand it the HDFS sink uses the
Your assumption is correct, as duplicates in a failure scenario will occur.
Thanks,
Rufus
On Tue, Sep 8, 2015 at 4:10 AM, Aljoscha Krettek
wrote:
> Hi,
> as I understand it the HDFS sink uses the transaction system to verify
> that all the elements in a transaction are written. This is what I w
No, you need the %M - else you'd still round down to the hour (since the
timestamp is rounded down, but they are still within the same hour)
Thanks,
Hari
On Tue, Sep 8, 2015 at 10:19 AM, IT CTO wrote:
> If the path stays with %h what would be the folder name with round 15 unit
> minute?
>
> בת
If the path stays with %h what would be the folder name with round 15 unit
minute?
בתאריך יום ג׳, 8 בספט׳ 2015, 19:52 מאת Hari Shreedharan <
hshreedha...@cloudera.com>:
> You can use roundValue as 15 and roundUnit as Minute - so a new dir would
> get created every 15 mins.
>
>
> Thanks,
> Hari
>
You can use roundValue as 15 and roundUnit as Minute - so a new dir would
get created every 15 mins.
Thanks,
Hari
On Tue, Sep 8, 2015 at 9:42 AM, Sutanu Das wrote:
> Hi Community,
>
>
>
> Is it possible to create /write HDFS dirs./folders every 15 mins?
>
>
>
> Right now, we use /hdfs_sink_pat
Hi Community,
Is it possible to create /write HDFS dirs./folders every 15 mins?
Right now, we use /hdfs_sink_path/%Y/%m/%d/%H --- ie using every Hour, But, Is
it possible to write every 15 mins? - not every minute...
If this is not possible out of the box config, can DEV team please advise whic
Thanks, FLUME-2787 and FLUME-2788 reported.
I will also try getting into the WIKI and learn about your contribution
process to see if I can help with it.
Do you have "HOWTO" for documentation contribution? :-)
Eran
On Tue, Sep 8, 2015 at 6:44 PM Hari Shreedharan
wrote:
> Can you file separate ji
Can you file separate jiras for each one here:
https://issues.apache.org/jira/browse/FLUME
Thanks,
Hari
On Tue, Sep 8, 2015 at 8:41 AM, IT CTO wrote:
> Hi,
> I just started learning flume and using it to send events to our HDFS
> cluster and log them in elasticsearch.
> I am using the latest b
Hi,
I just started learning flume and using it to send events to our HDFS
cluster and log them in elasticsearch.
I am using the latest build
Flume 1.6.0
Source code repository: https://git-wip-us.apache.org/repos/asf/flume.git
Revision: 2561a23240a71ba20bf288c7c2cda88f443c2080
Compiled by hshreedh
Looking at the code, I guess this sink is a bit different and the
"serializer" property doesn't seem to be used.
I see two options:
Either configure hdfs.writeFormat with an implementation of
SequenceFileSerializerType so it uses your own implementation of
SequenceFileSerializer.
Or extend HDFSEv
Von:Gonzalo Herreros
An: user@flume.apache.org,
Datum: 08.09.2015 09:29
Betreff:Re: How to customize the key in a HDFS SequenceFile sink
Thanks for your prompt reply. May I ask you to give me some more details.
I'm a little confused as I've read that the "hdfs.serializer" param
Hi,
as I understand it the HDFS sink uses the transaction system to verify that
all the elements in a transaction are written. This is what I would call
at-least-once semantics.
My question is now what happens if the writing fails in the middle of
writing the elements in the transaction. When the
The way Gonzalo said is how we always do it too.
2015-09-08 15:29 GMT+08:00 Gonzalo Herreros :
> I would implement a custom serializer and configure it in the standard
> Hdfs sink.
> That way you control how you build the key for each event.
>
> Regards,
> Gonzalo
>
> On 8 September 2015 at 06:42
I would implement a custom serializer and configure it in the standard Hdfs
sink.
That way you control how you build the key for each event.
Regards,
Gonzalo
On 8 September 2015 at 06:42,
wrote:
>
> Hello,
>
> I'm using Flume's HDFS SequenceFile sink for writing data to HDFS. I'm
> looking for
16 matches
Mail list logo