DataMovementEvent is a construct defined for an Input/Output pair to
communicate with each other. The actual information being passed between the 2
is not understood by the framework except in that, it is a byte payload to be
handed off from the source to the destination. Users are not expected
Hi
This looks like a 3-vertex DAG. It could be possibly be a linear DAG such as
Map1 -> Map2 -> Reduce3 or a Join DAG where
Map1 -> Reduce3 and Map2 -> Reduce3.
If you can get the application logs from YARN ( using bin/yarn logs
-applicationId application_1404180111945_438880 ), you will be a
se explain how the
> number of tasks got reduced here ?
>
> Thanks.
>
>
> On Thu, Jul 31, 2014 at 9:20 PM, Hitesh Shah wrote:
> Hi
>
> This looks like a 3-vertex DAG. It could be possibly be a linear DAG such as
> Map1 -> Map2 -> Reduce3 or a Join DAG where
>
However, the exception stack trace is misleading with respect to the actual
underlying problem. @Thad, could you file a bug in that regard?
thanks
— Hitesh
On Aug 7, 2014, at 1:27 PM, Bikas Saha wrote:
> The processor is expected to call start() before using the input. That’s the
> recommend
Hi Thad,
To clarify, are you looking for an API in the client to invoke a callback when
a session has shut down? Or invoke user logic in the AM when the session is
about to shut down?
For the former, given that client does not poll the session continuously,
detecting the session shutdown is o
It seems like the AM is binding to the external/public hostname and not the
internal IP.
Could you look for this log message in the AM logs: "Instantiated
DAGClientRPCServer at”. This will provide some information as to what the AM is
binding to.
thanks
— Hitesh
On Aug 11, 2014, at 7:43 AM,
Subroto, could you file a jira for this and mark it as a blocker for 0.5. This
should ideally be addressed before then. As commented on TEZ-1127, it is a
question as to what the default should be - whether HADOOP_COMMON_HOME or
HADOOP_PREFIX and to some extent, it needs to handle Windows deploym
Hello Robert,
Tez has gone through quite a few API changes since the 0.4.0 release. The
changes were mainly to clean up
and simplify the APIs to make it easier for users to write native Tez DAGs.
This had the unfortunate side-effect of making 0.5 incompatible
with Hive 0.13.x.
For now, I bel
Hi Robert,
From a Tez point of view, a user of the Tez APIs can define the container size
on a per vertex basis. I believe currently, Hive, when using Tez, uses a single
size for all its vertices.
thanks
— Hitesh
On Sep 9, 2014, at 6:37 PM, Grandl Robert wrote:
> Hi guys,
>
> It seems the
Hi Chris,
Unlike MR and its support of distributed cache, Tez does not make any
inferences into the structure of the LocalResources specified ( i.e structure
of tarball, jar, etc ) and therefore expects the user to modify the class path
as needed.
It might be something worth considering as a
Hello Kostas,
0.5.0 is released and should be available on the apache releases maven repo.
There should likely be a 0.5.1 release at some point in the near future with
various bug fixes.
As for 0.6.0-SNAPSHOT, we are adding features and bug fixes to this but it
should be compatible with 0.5.0.
Hi Robert,
This is a bug ( which has been seen in other scenarios too ). You can follow
some of the discussion related to this issue at
https://issues.apache.org/jira/browse/TEZ-1522.
thanks
— Hitesh
On Sep 12, 2014, at 11:20 AM, Grandl Robert wrote:
> Hi guys,
>
> During some of my expe
Hi Johannes,
To clarify, are you concerned about the 2 processes with pids 26964 and 27046
? If yes, the first one is the bash script invoked by the YARN NodeManager to
launch the actual TezChild JVM.
— HItesh
On Oct 13, 2014, at 3:18 AM, Johannes Zillmann wrote:
> Hey Bikas,
>
> this was
Hi all,
I am very pleased to announce that Jeff Zhang has been voted in as a committer
in
the Apache Tez project. We appreciate all the work Jeff has put into the
project so
far, and are looking forward to his future contributions.
Welcome aboard, Jeff.
Thanks,
— Hitesh on behalf of the Apache
Hi folks
As part of the Tez APIs, every object ( Input / Output / Processor /
EdgeManagerPlugin / VertexManagerPlugin ) can be associated with its own user
provided payload to set itself up. The format of this payload is not known to
Tez as it could be a java serialized object/protobuf/xml, et
Answers inline.
— Hitesh
On Oct 29, 2014, at 2:33 AM, Fabio wrote:
> Thanks Bikas for your answer and suggestion, actually my work deals more with
> high level modeling/behavior/performance of Tez, but there is another guy who
> is goign to handle Tez sources, I will suggest him to contribute
Hi
For the most part, each Hive CLI session or JDBC/ODBC connection to HiveServer2
would map to a single Application Master. HiveServer does have some
optimizations though ( to avoid the overhead cost of launching a new AM ) where
it tries to keep a pool of ApplicationMasters around and does s
+1(binding)
Verified sigs, checksums.
Built from source and ran example jobs on a single node cluster.
— Hitesh
On Nov 4, 2014, at 6:42 PM, Bikas Saha wrote:
>
> Folks,
>
> I have created an Apache Tez 0.5.2 release candidate (rc0).
>
> This is an incremental release that contains many
Hi Dharmesh
Would you mind filing a jira and attaching the application logs obtained by
running “bin/yarn logs -applicationId ”
Also, I am assuming you are setting “tez.am.log.level” to DEBUG to get debug
logging?
Also, when you query the timeline server, what are the calls that you are
maki
I am only looking at the TLS web UI. Planning to do programmatic
> > access if this runs.
> >
> > Thanks,
> > Dharmesh
> >
> > On Tue, Nov 11, 2014 at 9:41 PM, Hitesh Shah wrote:
> >>
> >> Hi Dharmesh
> >>
> >> Would you m
.HttpConnection.handle(HttpConnection.java:404)
> at
> org.mortbay.io.nio.SelectChannelEndPoint.run(SelectChannelEndPoint.java:410)
> at
> org.mortbay.thread.QueuedThreadPool$PoolThread.run(QueuedThreadPool.java:582)
>
> Nov 13, 2014 1:17:27 AM
> com.sun.jerse
Hello Fabio
We do not have a job trace file generated by Tez and therefore no simulator
that can re-run the trace. We do store some historical data for the job but the
level of tooling around it is pretty minimal.
— Hitesh
On Nov 14, 2014, at 3:29 AM, Fabio wrote:
> With SLS (Yarn Schedule
> Fabio
>
> On 11/14/2014 07:54 PM, Hitesh Shah wrote:
>> Hello Fabio
>>
>> We do not have a job trace file generated by Tez and therefore no simulator
>> that can re-run the trace. We do store some historical data for the job but
>> the level of tooling a
minutes job (not for me, for sure).
> I'm not even sure it's a job it is worth doing, since I wasn't able to find
> any information about the reliability of SLS as a Yarn simulator. Officially,
> it is just used to test the scheduler behavior, and as far as I remember it
>
Hi Subroto,
It could be an installer/distro issue. You may need to redirect your question
to the Hortonworks forums to get an answer on this.
But yes, you are right - the env needs to be setup correctly to load the
required native libs. Out of curiosity, are you using Ambari to setup your
cl
BCC’ed user@tez.
This question belongs to either the hive user list or the Hortonworks user
forums.
thanks
— Hitesh
On Dec 2, 2014, at 1:28 PM, Pala M Muthaia wrote:
> Hi,
>
> I am trying to get Tez sessions enabled with HS2. I start the HiveServer2
> instance with the flag "-hiveconf hive
In light of https://issues.apache.org/jira/browse/TEZ-1818, it might be better
to roll out a new RC with this fix?
Thoughts?
— Hitesh
On Dec 2, 2014, at 1:08 PM, Bikas Saha wrote:
> Folks,
>
>
>
> I have created an Apache Tez 0.5.3 release candidate (rc0).
>
>
>
> This is an incrementa
Hello folks
Could folks who are following these mailing lists do a raise of hands on which
versions of Hadoop you are trying to run Tez on?
1) 2.2.x
2) 2.3.x
3) 2.4.x
4) 2.5.x
5) 2.6.x
Given that we are building out the UI and for the large part, the History UI (
i.e. the one which will disp
Hi Fabio,
Regarding the second container assignment, the critical aspect is
"reusedContainer=true”. It is re-using the container used for the parent
vertex’s task hence the priority is not relevant. In such cases, eventually the
priority 4 container will be released without being used.
If you
, Moore, Douglas
wrote:
> 2.4.x and beyond for the customers we have
>
> Sent from my iPhone
>
>> On Dec 4, 2014, at 12:36 PM, Hitesh Shah wrote:
>>
>> Hello folks
>>
>> Could folks who are following these mailing lists do a raise of hands on
>> wh
We probably need to fix the docs that refer to
"tez.am.container.session.delay-allocation-millis”. Can you point which doc you
are referring to? This setting was removed in 0.5.x in favor of the min/max
release timeouts. To achieve the same behavior as
tez.am.container.session.delay-allocation
e is… etc..
> If the Hadoop UI integration could be delivered with 2.2 that would be
> awesome!
>
> best
> Johannes
>
>> On 08 Dec 2014, at 19:02, Hitesh Shah wrote:
>>
>> Thanks for the info, Douglas. We will not be dropping support for 2.4 for
>> sure
>>>
>>> Publishing multiple jars - do you mean per hadoop version supported ?
>>> That'll likely be very difficult to use - what would the dependency be ?
>>> Will there be binary compatibility issues if someone builds against tez jars
>>> b
Hello Xiaoyong,
Could you shed more light on the problems you have been encountering ( and with
which version of the hadoop )? Some of the details on how to use YARN timeline
are documented here: http://tez.apache.org/tez_yarn_timeline.html - let me know
if that helps.
Let us also know what ve
s
> using Tez 0.6.0.
>
> The tar under /apps/tez/ is tez-0.6.0.2.2.0.0-1084.tar.gz
>
> Xiaoyong
>
>
> -Original Message-
> From: Hitesh Shah [mailto:hit...@apache.org]
> Sent: Saturday, December 13, 2014 12:17 AM
> To: user@tez.apache.o
ar under /apps/tez/ is tez-0.6.0.2.2.0.0-1084.tar.gz
>
> Xiaoyong
>
>
> -Original Message-
> From: Hitesh Shah [mailto:hit...@apache.org]
> Sent: Saturday, December 13, 2014 12:17 AM
> To: user@tez.apache.org
> Subject: Re: Tez UI early trial?
>
> Hello Xiao
Hello again,
Wrong list for this question. Please send your hive specific question to the
hive-user list ( u...@hive.apache.org ).
thanks
— Hitesh
On Jan 5, 2015, at 2:23 PM, SkaterXu wrote:
> is there way to set map and reducer number for hive on tez manually?
> i want hive on tez to use a
Hello Peter
For the history logging exception, would you mind filing a jira for this issue
( https://issues.apache.org/jira/browse/TEZ ) and attach the yarn applications
log obtained by running the “yarn logs -applicationId” command?
The history logging happens on a separate thread and its ex
Hello
Thanks for tracking down the issue to the vcores setting. Let me dig into that.
Some initial questions:
- do you know if YARN has been configured to schedule on both memory and
vcores i.e using the DominantResourceScheduler?
- I am assuming that the max vcores per container is 1 but
l loop, it will
match against any pending task. Sorting by next schedule time means the thread
wakes up when it is time to run the next loop for a given container. Sorting
by expiry would imply scanning the whole list to find all containers whose
schedule time has elapsed.
>
>
Hi Johannes,
Sorry - missed this email earlier. I have seen this a couple of times but have
not been able to track down the root cause of this. Would you mind filing a
jira for this with the logs from an AM where you observed this? Also, if you
have a higher frequency of being able to reproduc
Hello Robert,
I believe currently the only way to do is this from the hive explain plan as
most of this information is inaccessible to Tez ( it is passed in via the user
payloads and only parsed by the Hive Processor ). I believe if you enable the
ATSHooks in Hive, it publishes such informatio
To clarify, it will be in one of the log-dirs ( not the working dir ) of the
AM’s container.
If you can access the AM logs from the RM UI, you can try out using
“http://people.apache.org/~gopalv/dagviz/“
When using “bin/yarn logs -applicationId ”, look for the files of
container_…._01 wh
Hi Robert,
Could you provide more details on the kind of Tez job you are running?
For example, if I run orderedwordcount, I see the HDFS counters show up
correctly:
2015-02-07 18:38:54,617 INFO [Dispatcher thread: Central]
history.HistoryEventHandler:
[HISTORY][DAG:dag_1423172314047_0008_1
Just to be clear, if you are looking for AM debug logs, you can enable debug
logging per app. Just use “word count -Dtez.am.log.level=DEBUG input output”.
— Hitesh
On Feb 10, 2015, at 4:02 AM, Fabio C. wrote:
> Hi everyone,
> I was running the tez wordcount example on a 6 nodes cluster. The i
Thanks for the feedback, Johannes.
Would it be possible for you to file a jira for the performance issue that you
are seeing with logs? Please strip out necessary data to hide the customer
info, etc. The logs that would be most useful are:
- comparison logs of an MR job vs a Tez job showing
A clarification for (2), you can share an AM across multiple users by using
form of proxy users and passing in the required delegation tokens to talk to
various services such as HDFS. Also, HiveServer2 when the doAs mode is set to
false, runs all AMs as user hive but can effectively run queries
Hello Alexander,
Are you using Tez natively or via Hive/Pig/Cascading, etc?
To a large extent, most users I have encountered tend to have
tez.am.resource.memory.mb sized to be between 4-8 GB though in some cases, (
until TEZ-776 is addressed ), this might need to increased for DAGs which have
Hello Johannes,
This is something we have discussed quite often but have not got around to
implementing this. There might be an open jira related to “pipelining” of
splits. If you cannot find it, please go ahead and create one.
The general issues with these are:
- how to handle dynamic crea
Hello
First issue from the stack trace:
"org.apache.tez.dag.api.TezUncheckedException: Invalid configuration of tez
jars, tez.lib.uris is not defined in the configuration”. It looks like your
first run failed due to this. The second run seemed to be configured correctly
but failed for a diffe
ter error was
> raised...
>
>
> regards
>
>
> ----
> On Tue, 3/17/15, Hitesh Shah wrote:
>
> Subject: Re: error after installation - TezSession has already shutdown
> To: user@tez.apache.org
> Received: Tuesday
Hi Michael,
Yes - that is indeed an error regarding the version of protobuf that needs to
be used. Pretty much all of the Hadoop ecosystem components have standardized
on protobuf-2.5.0. Newer versions would likely have worked assuming protobuf
retained compatibility.
In any case, you do need
Hi Chris,
I don’t believe this issue has been seen before. Could you file a jira for this
with the full application logs ( obtained via bin/yarn logs -application ) and
the configuration used?
thanks
— Hitesh
On Mar 23, 2015, at 1:01 PM, Chris K Wensel wrote:
> Hey all
>
> We have a user r
Hi
I just downloaded the src-tar for the 0.6.0 release and built it without any
issues. At times, I have not seen this but others have faced issues at times (
based on general searches[1] ) where there have been temporal issues
downloading node locally at times.
[1] https://github.com/eirslett
timeline server and tomcat.
>
> and I also changed script/config.js to the correct timeline URL and RMUrl (my
> timeline webapp port is 50040, RM webapp port is 50030)
>
> then I open http://x.x.x.x:8080/tez-ui-xx/index.html, which is a blank page.
>
> I am strictly refer to
Thanks for the heads up, Azuryy. I will pass the information along to the
relevant folks at Hortonworks. And yes, you are right. At some point, we merged
the code into the common TezClient so that it became easier for the user to
toggle session-mode on/off via configuration and not have to write
Not much of help but you can start with these:
http://tez.apache.org/releases/0.6.0/tez-api-javadocs/org/apache/tez/dag/api/TezConfiguration.html
http://tez.apache.org/releases/0.6.0/tez-runtime-library-javadocs/org/apache/tez/runtime/library/api/TezRuntimeConfiguration.html
The above 2 files’ d
Hi Johannes
Not sure if anyone has seen this earlier. Do you know if the machines have
enough memory to run the no. of tasks/containers that you are launching? Also,
I am assuming that you are compiling and running against the same jdk version?
Would you mind sharing the details on what java v
Hello Xiaoyong,
I believe you might get better help on the hive user mailing lists given that
templeton is part of hive. I am guessing that Templeton has some hooks to
recognize indirectly launched MR jobs but may not have built out the same
support to recognize tez jobs.
thanks
— Hitesh
O
@r7raul1984, would you mind filing a documentation jira for your question. The
list that Rajesh provided might be good to formalize into a doc and/or wiki.
Also, please take a look at https://issues.apache.org/jira/browse/TEZ-2294 to
see all the list of parameters. If you see something off or n
As a general point for other users, although this is not fully supported, you
can try running 0.5.3 with the ATS history logging enabled and then setup the
UI using the 0.6 build with the UI configured to point to the ATS server.
Almost all of the required data published to ATS should be in the
.0 and Tez is 0.53.
> I try http://192.168.117.117:8288/ws/v1/timeline/TEZ_DAG_ID?limit=11 will
> retrun result ;
> But Tez UI always display .
> <54AA3VC(8~{WI`O`(04-29-08-47-14).png>
>
> r7raul1...@163.com
>
> From: Hitesh Shah
> Date: 2015-04-27 23:19
>
This might be a mail that is better suited for the user@hive mailing list to
start with.
thanks
— Hitesh
On May 5, 2015, at 12:58 AM, r7raul1...@163.com wrote:
> I change the sql where condition to (where t.update_time >= '2015-05-04') ,
> the sql can return result for a while. Because t.up
t time. and at that time shuffle keep running for ever.
> which version of tez r u using?
>
>
> On 2015-05-06 06:02 , Hitesh Shah Wrote:
>
> This might be a mail that is better suited for the user@hive mailing list to
> start with.
>
> thanks
> — Hitesh
>
Hi all,
I am very pleased to announce that Sreenath Somarajapuram has been voted in as
a committer in the Apache Tez project. We appreciate all the work Sreenath has
put into the project so far ( especially for the Tez UI ), and are looking
forward to his future contributions.
Welcome aboard
This will require a change in Hive/Pig and a modification to the Tez UI.
For every vertex, there is a ProcessorDescriptor. This supports a
setHistoryText API.
/**
* Provide a human-readable version of the user payload that can be
* used in the TEZ UI
* @param historyText History text
* For
Hello Amit,
For the most part, all that local fetch does is that in the case where the
upstream vertex's output is on the same host where the downstream vertex task
is running, the fetcher reads the data directly from disk instead of going via
the http-based shuffle handler. This is an optimiz
+1 on dropping 1.6 support from 0.8.0 onwards.
@Mohammad, Hadoop is dropping support for 1.6 from 2.7.0 onwards. I am guessing
other ecosystem projects will soon follow suit.
— Hitesh
On May 15, 2015, at 2:16 PM, Mohammad Islam wrote:
> Hi Sid,
> What are the statuses of other Hadoop project
.
thanks
— Hitesh
On May 16, 2015, at 2:15 AM, Xiaoyong Zhu wrote:
> I see, thanks Hitesh!
>
> That being said - currently Hive does not use ProcessorDescriptor to log the
> query info, right?
>
> Xiaoyong
>
> -Original Message-
> From: Hitesh Shah [mailto
:
> +1 non-binding on dropping 1.6. However given that we compile against
> hadoop2.4 and hadoop2.2 which has requireJavaVersion 1.6 will it be an issue?
>
>
>
>
> On 5/16/15, 3:54 AM, "Hitesh Shah" wrote:
>
>> +1 on dropping 1.6 support from 0.8
There is nothing that prevents a processor running and finishing without even
reading any data from any input. The only point when the processor blocks is
when it tries to read data from a particular input that has not yet finished
fetching all of its data.
That said, a processor cannot yet que
This is with respect to how work is assigned to a Task. For a shuffle edge, a
Task’s input is determined based on the partitions and how partitions are
assigned to a Task. For a vertex reading data from HDFS ( initial input ). this
is effectively random as the input data is split up and then ass
ns don¹t matter, what matters
> are downstream sourceVersions of tools which use Tez.
>
> Cheers,
> Gopal
>
> On 5/16/15, 12:07 PM, "Hitesh Shah" wrote:
>
>> Excellent point @Prakash. I would probably change my vote to a -0 on that
>> point itself.
>>
Hello Andre,
Could you file a JIRA for this and upload the logs around the point where it
hangs?
thanks
— Hitesh
On May 21, 2015, at 7:55 AM, Andre Kelpe wrote:
> Hi,
>
> we have a big test suite for lingual, our SQL layer for cascading. We are
> trying very hard to make it work correctly
There is some history logging done that can be enabled via the
SimpleHistoryLogger. This activates by default if ATS logger is not enabled.
This is not fully compatible with the ATS data and also as it is mostly
experimental, it may not have all the data. To use it, you can configure the
“tez.h
Hello Patcharee
Could you start with sending a mail to users@pig to see if they have come
across this issue first? Also, can you check the application master logs to see
if there are any errors ( might be useful to enable DEBUG level logging to get
more information )?
thanks
— Hitesh
On May
In hadoop 2.4, YARN timeline does not really support proper security. It does
not have any ACL support ( implemented using something called domains ).
The property that you mentioned needs to be set as it is a form of a warning to
the user that you are running Tez with YARN Timeline with ACLs e
These are hive-specific questions. Could you please send this mail to the
user@hive list instead?
thanks
— Hitesh
On May 26, 2015, at 10:28 PM, r7raul1...@163.com wrote:
> How to find query log when I use hiveserver2 on tez? My enviroment is hive
> 1.1.0+tez0.53 use beeline to connect hiveser
To clarify, given that the error is showing up with
container_1432885077153_0004_01_05, that means that the AM launched
properly.
Use “bin/yarn logs -applicationId application_1432885077153_0004" to get the
logs. See if there are any errors for the logs for
container_1432885077153_0004_01
Hi all,
I am very pleased to announce that Jeff Zhang has been voted in as a member of
the Apache Tez PMC. We appreciate all the work Jeff has put into the project so
far, and are looking forward to his future (greater) contributions.
Please join me in welcoming Jeff to the Tez PMC. Congratul
If you compiled Tez against hadoop-2.6.0 and are deploying it on a hadoop-2.5.0
cluster, you should disable tez acls as YARN timeline in 2.5.0 does not support
ACLs.
Please set tez.am.acls.enabled to false as the Timeline layer is trying to
enforce acls for the history data.
thanks
— Hitesh
O
There might be a better way to approach this.
You can add the following to the top-level pom.xml in the repositories section:
cloudera-repo
Cloudera Repository
https://repository.cloudera.com/artifactory/cloudera-repos
true
false
Templeton has had some bugs related to not adding tez-site.xml to the launcher
task from which the actual hive job runs. I think they are fixed in the newer
versions. Can you try setting the tez ats property into webhcat-site or
hive-site and see if it propagates all the way to the MR task which
Unless you can pinpoint the problem to something Tez specific, hive-specific
questions might be better off being asked on user@hive initially as there is a
larger group there that understands Hive as compared to the Tez community.
FWIW, TezTask error 1 means “something in the Hive layer using T
This may likely be TaskAttemptListenerImpTezDag used for communication between
the AM and the task containers. There are no configuration knobs for it
currently as it is always internal to a cluster.
If you don’t mind, can you file a jira and contribute a patch to add a
clarifying log message t
That particular log is a red herring and not really an issue that is causing
the failure.
The main problem based on the log is this:
2015-06-17 18:00:43,543 INFO [AsyncDispatcher event handler]
history.HistoryEventHandler:
[HISTORY][DAG:dag_1433219182593_180456_3][Event:DAG_FINISHED]:
dagId
nostics=[Vertex received Kill while in RUNNING state., Vertex killed as
> other vertex failed. failedTasks:0, Vertex vertex_1433219182593_180456_3_00
> [Map 1] killed/failed due to:null]
> DAG failed due to vertex failure. failedVertices:1 killedVertices:1
> FAILED: Execution Error
null]
> DAG failed due to vertex failure. failedVertices:1 killedVertices:1
> FAILED: Execution Error, return code 2 from
> org.apache.hadoop.hive.ql.exec.tez.TezTask
>
> I think maybe first successfule job delete tez-conf.pb ?
>
> r7raul1...@163.com
>
> From:
The build tries to download node and then install it. Did the download fail for
some reason? You can try running in debug mode i.e. “mvn -X” to see why it
could be failing.
— Hitesh
On Jun 19, 2015, at 8:17 AM, amit kumar wrote:
> Hi,
>
>
> I am trying to build tez-0.7.0, but can not build
Yes something along those lines.
This might help a bit more: http://techtonka.com/?p=174
thanks
— Hitesh
On Jun 18, 2015, at 6:44 PM, r7raul1...@163.com wrote:
>
> log like this hdfs-audit.log.9 ?
> r7raul1...@163.com
>
> From: Hitesh Shah
> Date: 2015-06-19 02:28
> T
o if there is any location from where I can download the jar
> (artifact) directly instead of building it.
>
> It would be very helpful.
>
> Thanks.
>
>
> On Fri, Jun 19, 2015 at 3:31 PM, Hitesh Shah wrote:
> The build tries to download node and then install it.
The error seems to indicate that the create dir/write to HDFS failed. Can you
check compare what user you are running as and whether the user has permissions
to create/write to the directories in the path below? Furthermore, you may wish
to check if the datanodes are alive and also look for erro
A couple of things to do:
1) (optional) For tez.lib.uris, set it to "${fs.defaultFS}/apps/tez/tez.tar.gz”
- this tez.tar.gz should come from tez-dist/target/. Given that your basic job
is working, you can ignore this for now but it is the recommended way to deploy
tez.
2) On a fresh setup, do
Hello folks,
The Apache Tez team is proud to announce the release of Apache Tez version
0.5.4
The Apache Tez project is aimed at creating a framework to build efficient and
scalable data processing applications that can be modeled as data flow graphs.
This release is a maintenance release fo
Move to user@hive. BCC’ed user@tez.
— Hitesh
On Jun 30, 2015, at 1:44 AM, patcharee wrote:
> Hi,
>
> I am using hive 0.14 + tez 0.5. It fails to alter table concatenate
> occasionally (see the exception below). It is strange that it fails from time
> to time not predictable. However, it wor
From a Tez perspective, there was a major compatibility change between Tez 0.4
and Tez 0.5. However, Tez-0.7.x and Tez-0.6.x are compatible with Tez-0.5.x.
I believe Hive 0.13 is compatible only with Tez 0.4.
For Hive 0.14 onwards ( including the Hive-1.x. releases ), they should work
with any
For data skew, you may also want to consider enabling
“tez.task.generate.counters.per.io”. This enables counters on a per edge basis
which is more helpful for complex DAGs.
— Hitesh
On Jul 8, 2015, at 10:29 PM, Joe Zhang (SDE) wrote:
> Hi Rajesh:
>
> Thanks for your reply. I want to know mo
Thanks for digging into the issue, Rajat.
Mind filing a jira for this and uploading your patch ( without the frontend
plugin change)?
thanks
— Hitesh
On Jul 15, 2015, at 3:53 PM, Rajat Jain wrote:
> I run this command:
>
> mvn clean install -DskipTests
>
> It automatically picks up Hadoo
I think this might be a bug somewhere in YARN ( specifically the fair scheduler
layer ) as Tez has no control over the data being displayed. The resource usage
and accounting is all handled within the ResourceManager in YARN.
You should probably file a jira against YARN for this issue and attac
Hi Sachin,
Quite a few folks have reported seeing this and there is an open jira for this.
From a previous reply by Prakash:
{quote}
This is a known bug. TEZ-2560 is tracking it – will put in a fix soon.
For now you can use a maven version < 3.3 OR change the frontend-maven-plugin
version to
1 - 100 of 216 matches
Mail list logo