hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sushanth Sowmyan <khorg...@gmail.com>
Subject Re: Apache Hive 0.13.1
Date Thu, 08 May 2014 02:22:16 GMT
After much experimentation with git bisect (which is very powerful),
I've narrowed down the test failures reported yesterday. The failures
are appearing from the following:

HIVE-6945:
org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_nullformatCTAS
org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_show_create_table_alter
org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_show_tblproperties
org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_unset_table_view_property
org.apache.hadoop.hive.cli.TestNegativeCliDriver.testNegativeCliDriver_unset_table_property

HIVE-6846:
org.apache.hive.service.cli.TestScratchDir.testLocalScratchDirs

HIVE-6826:
org.apache.hadoop.hive.cli.TestMinimrCliDriver.testCliDriver_bucketmapjoin6


Of the above, the second jira was already in 0.13.0. I'll comment up
on those jiras asking the committers involved in those bugs and to
help debug the issue. If anyone is interested in the git bisect logs
for these, they're available on
http://people.apache.org/~khorgath/releases/0.13.1_RC0/test_failures/


On Tue, May 6, 2014 at 6:41 PM, Sushanth Sowmyan <khorgath@gmail.com> wrote:
> Also, I wanted to throw in one more bit for those of you that are
> interested in tinkering along :
>
> http://people.apache.org/~khorgath/releases/0.13.1_RC0/relprep.pl
> http://people.apache.org/~khorgath/releases/0.13.1_RC0/requested_jiras
>
> This is the script and config file I'm using to generate this release.
>
> It's very much a hack right now, and I hope to improve it to
> streamline releases in the future, but how it can be used right now is
> this way:
>
> a) Put it in a hive git repo (and not have any changes that have not
> been committed - this script will checkout a new branch and commit
> things to that branch, so you want to make sure to have a clean repo)
> b) Put the file requested_jiras in that dir as well.
> c) Run the script from there.
>
> It will check the differences between the branch being released
> ("branch-0.13" is hardcoded currently as a global), and looks at all
> the commit logs in trunk that correspond to the jiras requested in the
> requested_jiras file, sorts them in the order they were committed, and
> then checks out a new branch called relprep-branch-0.13-<timestamp>,
> and attempts to cherry-pick those commits in.
>
> For some patches, this will not work, so there is an override
> mechanism provided by entries in the requested_jiras file, as can be
> observed in the file I mention above.
>
> At the end of it, you'll have your 0.13.1 repo reproduction to test
> against if you so desire.
>
> Known Bugs :
>
> a) I use system("....") or die "...";, which is faulty in that the
> "die" code will never be reached. I need to fix this, but all the
> system calls were working for me, and I'd much rather focus on the
> release now, and improve this script later. This is a TODO
> b) Some patches (those generated with --no-prefix) don't work with
> older versions of git. You'll need a 1.8.x git for them, or you have
> to generate git patches without --no-prefix.
>
>
>
> On Tue, May 6, 2014 at 6:21 PM, Sushanth Sowmyan <khorgath@gmail.com> wrote:
>> Hi Folks,
>>
>> After a run of the ptest framework across the 0.13.1 codebase, we have
>> a couple of test failures that I'm trying to track down and resolve.
>>
>> If any of you are interested in looking at it on your own in the
>> meanwhile, the conglomerate patch of all the patches I'm forward
>> porting into 0.13.1 is over at
>> http://people.apache.org/~khorgath/releases/0.13.1_RC0/0.13.1.gdiff.patch
>>
>> The current tests that are failing are as follows:
>>
>> org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_nullformatCTAS
>> org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_show_create_table_alter
>> org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_show_tblproperties
>> org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver_unset_table_view_property
>> org.apache.hadoop.hive.cli.TestNegativeCliDriver.testNegativeCliDriver_unset_table_property
>> org.apache.hive.service.cli.TestScratchDir.testLocalScratchDirs
>>
>> I'll update and follow up with patch devs as and when I find out the
>> source for these errors.
>>
>> Thanks,
>> -Sushanth
>>
>> On Mon, May 5, 2014 at 6:26 PM, Sushanth Sowmyan <khorgath@gmail.com> wrote:
>>> Hi Folks,
>>>
>>> It's past 6pm PDT on May 5th 2014, so I'm beginning the process to
>>> generate the 0.13.1 RC0.
>>>
>>> I've received backport patches for 0.13 for some of the jiras that did
>>> not cleanly apply back from 0.14 to 0.13, and I'm thankful for them.
>>> All requested patches are cleanly applying and compiling. Tests are
>>> still running, and I expect them to run for a while more still - I
>>> will make sure to wait for them to complete before actually cutting a
>>> release candidate.
>>>
>>> The very first thing I'm supposed to do per the HowtoRelease wiki, is
>>> to unlink any open non-blockers from the release, and then send out a
>>> mail to the dev list confirming that there are no open non-blockers
>>> for this release version. So, this is that mail:
>>>
>>>
>>> The jira query string I used to verify this is as follows:
>>>
>>> project = HIVE AND priority not in (Critical, Major, Minor, Trivial)
>>> AND fixVersion = 0.13.1 AND status = Open
>>>
>>> The good news is that the 0.13.1 branch name was only created a couple
>>> of hours back, so no issues have been created on this, so there were
>>> none to unlink. :)
>>>
>>>
>>> On Fri, May 2, 2014 at 6:22 PM, Sushanth Sowmyan <khorgath@gmail.com> wrote:
>>>> Hi Folks,
>>>>
>>>> It is now past 6pm PDT May 2nd 2014, so I have locked down the table
>>>> of jiras for inclusion into 0.13.1. The jiras included are as listed
>>>> in :
>>>>
>>>> https://cwiki.apache.org/confluence/display/Hive/Hive+0.13.1+Release+tracking
>>>>
>>>> The jiras being included are :
>>>>
>>>> HIVE-6919, HIVE-6921, HIVE-6957, HIVE-6945, HIVE-6862,
>>>> HIVE-6883, HIVE-6961, HIVE-4576, HIVE-6952, HIVE-6828
>>>> HIVE-6898, HIVE-6783, HIVE-6955, HIVE-6877, HIVE-6985
>>>> HIVE-7001
>>>>
>>>> The inclusion requestors for the above are :
>>>>
>>>> thejas, ashutoshc, sershe, ekoifman, prasanth_j, hagleitn, szehon,
>>>> rhbutani and amareshwari.
>>>>
>>>> If you are in this list, I might be pinging you in the next couple of
>>>> days for help if the patch does not cleanly apply, or causes
>>>> additional failures, and I appreciate any help you can tend.
>>>>
>>>> There is one jira on that list, HIVE-6826, which is in the list, and
>>>> will still be included if it moves from Patch-available to Committed
>>>> in 0.14 by Monday when I roll out the RC, which I have assurances from
>>>> vikram that it will, given that it's waiting on a test result.
>>>>
>>>> If anyone has any further breaking jiras that they absolutely want in
>>>> this release, please do still bring it up, I've created a new table
>>>> for them in the release tracking wiki page. I do, however, raise the
>>>> bar for inclusion by asking two other committers to also +1 such an
>>>> inclusion request at this stage.
>>>>
>>>> Hopefully, we'll see an RC and a release pretty soon! :)
>>>>
>>>> Thanks all!
>>>> -Sushanth
>>>>
>>>>
>>>> On Thu, May 1, 2014 at 2:34 PM, Sushanth Sowmyan <khorgath@gmail.com>
wrote:
>>>>> Hi All,
>>>>>
>>>>> A quick reminder - per the timeline discussed so far, the list of
>>>>> jiras for inclusion in 0.13.1 RC0 freezes tomorrow (May 2nd) evening
>>>>> at 6pm PDT. Please make sure you ask for inclusion or edit the wiki
>>>>> with any further requests so that we can get this release done in a
>>>>> timely manner.
>>>>>
>>>>> Thanks,
>>>>> -Sushanth
>>>>>
>>>>>
>>>>> On Mon, Apr 28, 2014 at 1:39 PM, Sushanth Sowmyan <khorgath@gmail.com>
wrote:
>>>>>> Good question - I think that if the number of patches is small (
< ~20
>>>>>> which I expect because this is a stabilization release soon after
>>>>>> 0.13.0), I'd prefer to commit them myself, so that I can retain order
>>>>>> of patch application and make sure they all apply cleanly. I'm also
>>>>>> actually using this as an opportunity towards testing out an
>>>>>> semi-automated script that reads the requested jiras to do this -
I
>>>>>> figure it's useful to test something like that out on a smaller
>>>>>> release that'll make further releases easier.
>>>>>>
>>>>>> On Mon, Apr 28, 2014 at 11:42 AM, Thejas Nair <thejas@hortonworks.com>
wrote:
>>>>>>> Sushanth,
>>>>>>> How do you want to go about committing the patches into 0.13
branch ?
>>>>>>> Will you be able to commit the requested patches into 0.13.1
yourself, or
>>>>>>> do you want requestors to discuss it with you and then commit
the patches
>>>>>>> themselves (if the requestor is a committer).
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Thejas
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On Sat, Apr 26, 2014 at 2:19 PM, Sushanth Sowmyan <khorgath@gmail.com>wrote:
>>>>>>>
>>>>>>>> Added.
>>>>>>>>
>>>>>>>> If others have difficulty editing( I can't figure out how
to change editing
>>>>>>>> privileges, but it seems to indicate that others can edit)
the page, I'll
>>>>>>>> accept replies to this thread as well and can add it in.
>>>>>>>>  On Apr 25, 2014 6:25 PM, "Sergey Shelukhin" <sergey@hortonworks.com>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>> > I don't have access to edit this page (or cannot figure
out the UI).
>>>>>>>> > Username sershe.
>>>>>>>> > Can you add
>>>>>>>> > HIVE-6961 : Drop partitions treats partition columns
as strings (area -
>>>>>>>> > metastore)
>>>>>>>> >
>>>>>>>> >
>>>>>>>> > On Fri, Apr 25, 2014 at 4:20 PM, Sushanth Sowmyan <khorgath@gmail.com
>>>>>>>> > >wrote:
>>>>>>>> >
>>>>>>>> > > I've created the following wiki link :
>>>>>>>> > >
>>>>>>>> > >
>>>>>>>> >
>>>>>>>> https://cwiki.apache.org/confluence/display/Hive/Hive+0.13.1+Release+tracking
>>>>>>>> > >
>>>>>>>> > > People should be able to request additional jiras
by adding it to the
>>>>>>>> > > list. I think it might make sense to halt addition
of requests to the
>>>>>>>> > > list 3 days before the RC is cut, so as to prevent
an endless-tail
>>>>>>>> > > scenario, unless the bug in question is a breaking
severe issue,
>>>>>>>> > > where, yes, after discussion, we can vote to add
it to the list. That
>>>>>>>> > > also gives us time to run a full suite of tests
on a stable build
>>>>>>>> > > before we cut the RC.
>>>>>>>> > >
>>>>>>>> > > I propose that the first RC (RC0) be built on Monday
May 5th at 6pm
>>>>>>>> > > PDT, and the jira list on the wiki be closed to
open/easy additions at
>>>>>>>> > > 6pm PDT on Friday May 2nd.
>>>>>>>> > >
>>>>>>>> > >
>>>>>>>> > > On Fri, Apr 25, 2014 at 2:40 PM, Gunther Hagleitner
>>>>>>>> > > <ghagleitner@hortonworks.com> wrote:
>>>>>>>> > > > Sorry - HIVE-6824 isn't needed. Just the other
3. My bad.
>>>>>>>> > > >
>>>>>>>> > > > Thanks,
>>>>>>>> > > > Gunther.
>>>>>>>> > > >
>>>>>>>> > > >
>>>>>>>> > > > On Fri, Apr 25, 2014 at 2:10 PM, Gunther Hagleitner
<
>>>>>>>> > > > ghagleitner@hortonworks.com> wrote:
>>>>>>>> > > >
>>>>>>>> > > >> I'd like to request to include these Tez
fixes:
>>>>>>>> > > >>
>>>>>>>> > > >> HIVE-6824, HIVE-6826, HIVE-6828, HIVE-6898
>>>>>>>> > > >>
>>>>>>>> > > >> Thanks,
>>>>>>>> > > >> Gunther.
>>>>>>>> > > >>
>>>>>>>> > > >>
>>>>>>>> > > >> On Fri, Apr 25, 2014 at 11:59 AM, Sushanth
Sowmyan <
>>>>>>>> > khorgath@gmail.com
>>>>>>>> > > >wrote:
>>>>>>>> > > >>
>>>>>>>> > > >>> True, I was counting two weeks from
today, but 0.13 has already
>>>>>>>> been
>>>>>>>> > > >>> out for a week. I'm amenable to having
an RC1 out on May 5th. If
>>>>>>>> any
>>>>>>>> > > >>> further issues appear that block,
then we can deal with them in an
>>>>>>>> > > >>> RC2/etc modification to that.
>>>>>>>> > > >>>
>>>>>>>> > > >>> On Fri, Apr 25, 2014 at 11:45 AM,
Thejas Nair <
>>>>>>>> > thejas@hortonworks.com>
>>>>>>>> > > >>> wrote:
>>>>>>>> > > >>> > On Fri, Apr 25, 2014 at 11:33
AM, Sushanth Sowmyan <
>>>>>>>> > > khorgath@gmail.com
>>>>>>>> > > >>> >wrote:
>>>>>>>> > > >>> >
>>>>>>>> > > >>> >>
>>>>>>>> > > >>> >> I think it's important to
get a bugfix/stabilization release
>>>>>>>> > > >>> >> reasonably quickly, but it's
also important to give people a
>>>>>>>> > little
>>>>>>>> > > >>> >> time to try out 0.13, discover/report
bugs and fix them. So I
>>>>>>>> > think
>>>>>>>> > > >>> >> about two weeks is a good
point? And instead of releasing an RC
>>>>>>>> > on a
>>>>>>>> > > >>> >> friday, I'm thinking of pushing
it out to Monday - does 12th May
>>>>>>>> > > sound
>>>>>>>> > > >>> >> good to everyone?
>>>>>>>> > > >>> >>
>>>>>>>> > > >>> >
>>>>>>>> > > >>> > I think we can aim for an earlier
date. Most of these issues seem
>>>>>>>> > to
>>>>>>>> > > be
>>>>>>>> > > >>> > already committed to trunk or
have patches available. So the
>>>>>>>> > > remaining
>>>>>>>> > > >>> ones
>>>>>>>> > > >>> > also might get committed to trunk
by early next week. How about
>>>>>>>> > > shooting
>>>>>>>> > > >>> > for May 5th (Monday) ?
>>>>>>>> > > >>> > By then 0.13 would also have
been out for 2 weeks. If we have any
>>>>>>>> > new
>>>>>>>> > > >>> > critical bug reported that needs
a fix, we can hold off on the RC
>>>>>>>> > for
>>>>>>>> > > >>> few
>>>>>>>> > > >>> > days.
>>>>>>>> > > >>> > What do you think ?
>>>>>>>> > > >>> >
>>>>>>>> > > >>> > Thanks,
>>>>>>>> > > >>> > Thejas
>>>>>>>> > > >>> >
>>>>>>>> > > >>> > --
>>>>>>>> > > >>> > CONFIDENTIALITY NOTICE
>>>>>>>> > > >>> > NOTICE: This message is intended
for the use of the individual or
>>>>>>>> > > >>> entity to
>>>>>>>> > > >>> > which it is addressed and may
contain information that is
>>>>>>>> > > confidential,
>>>>>>>> > > >>> > privileged and exempt from disclosure
under applicable law. If
>>>>>>>> the
>>>>>>>> > > >>> reader
>>>>>>>> > > >>> > of this message is not the intended
recipient, you are hereby
>>>>>>>> > > notified
>>>>>>>> > > >>> that
>>>>>>>> > > >>> > any printing, copying, dissemination,
distribution, disclosure or
>>>>>>>> > > >>> > forwarding of this communication
is strictly prohibited. If you
>>>>>>>> > have
>>>>>>>> > > >>> > received this communication in
error, please contact the sender
>>>>>>>> > > >>> immediately
>>>>>>>> > > >>> > and delete it from your system.
Thank You.
>>>>>>>> > > >>>
>>>>>>>> > > >>
>>>>>>>> > > >>
>>>>>>>> > > >
>>>>>>>> > > > --
>>>>>>>> > > > CONFIDENTIALITY NOTICE
>>>>>>>> > > > NOTICE: This message is intended for the use
of the individual or
>>>>>>>> > entity
>>>>>>>> > > to
>>>>>>>> > > > which it is addressed and may contain information
that is
>>>>>>>> confidential,
>>>>>>>> > > > privileged and exempt from disclosure under
applicable law. If the
>>>>>>>> > reader
>>>>>>>> > > > of this message is not the intended recipient,
you are hereby
>>>>>>>> notified
>>>>>>>> > > that
>>>>>>>> > > > any printing, copying, dissemination, distribution,
disclosure or
>>>>>>>> > > > forwarding of this communication is strictly
prohibited. If you have
>>>>>>>> > > > received this communication in error, please
contact the sender
>>>>>>>> > > immediately
>>>>>>>> > > > and delete it from your system. Thank You.
>>>>>>>> > >
>>>>>>>> >
>>>>>>>> > --
>>>>>>>> > CONFIDENTIALITY NOTICE
>>>>>>>> > NOTICE: This message is intended for the use of the
individual or entity
>>>>>>>> to
>>>>>>>> > which it is addressed and may contain information that
is confidential,
>>>>>>>> > privileged and exempt from disclosure under applicable
law. If the reader
>>>>>>>> > of this message is not the intended recipient, you are
hereby notified
>>>>>>>> that
>>>>>>>> > any printing, copying, dissemination, distribution,
disclosure or
>>>>>>>> > forwarding of this communication is strictly prohibited.
If you have
>>>>>>>> > received this communication in error, please contact
the sender
>>>>>>>> immediately
>>>>>>>> > and delete it from your system. Thank You.
>>>>>>>> >
>>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> CONFIDENTIALITY NOTICE
>>>>>>> NOTICE: This message is intended for the use of the individual
or entity to
>>>>>>> which it is addressed and may contain information that is confidential,
>>>>>>> privileged and exempt from disclosure under applicable law. If
the reader
>>>>>>> of this message is not the intended recipient, you are hereby
notified that
>>>>>>> any printing, copying, dissemination, distribution, disclosure
or
>>>>>>> forwarding of this communication is strictly prohibited. If you
have
>>>>>>> received this communication in error, please contact the sender
immediately
>>>>>>> and delete it from your system. Thank You.

Mime
View raw message