couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jan Lehnardt <...@apache.org>
Subject Re: Failing CLI tests
Date Tue, 04 Dec 2012 13:17:05 GMT

On Dec 3, 2012, at 03:36 , Noah Slater <nslater@apache.org> wrote:

> Okay, I have side-stepped the issue with:
> 
> export AM_DISTCHECK_CONFIGURE_FLAGS=--disable-tests
> 
> (--disable-tests was something I added to just skip all this stuff when I
> am working on the build...)
> 
> But my point remains.
> 
> I am using the largest AWS instance size, and I cannot make distcheck. We
> can't release without this, obviously.
> 
> Would really appreciate some help getting this fixed! :D

Curious indeed, it works fairly wellâ„¢ on my MacBook Air & Mac Mini on
SSD & spinning rust.

Do we have any AWS-savvy readers here that could aide the debugging of
this?

Cheers
Jan
-- 

> On 3 December 2012 02:18, Noah Slater <nslater@apache.org> wrote:
> 
>> Just tried another run, and it froze up in the tests. For 10 minutes or
>> more. So I stopped it and started again.
>> 
>> ok 10 batch_save
>> ok 11 bulk_docs
>> ok 12 changes
>> ok 13 coffee
>> ok 14 compact
>> ok 15 config
>> ok 16 conflicts
>> ok 17 content_negotiation
>> ok 18 cookie_auth
>> ok 19 copy_doc
>> 
>> 
>> 
>> ^Cmake[1]: *** [check-js] Interrupt
>> make: *** [distcheck] Interrupt
>> 
>> And it happens on a second run...
>> 
>> ok 37 oauth_users_db
>> ok 38 proxyauth
>> ok 39 purge
>> ok 40 reader_acl
>> ok 41 recreate_doc
>> ok 42 reduce
>> ok 43 reduce_builtin
>> ok 44 reduce_false
>> ok 45 reduce_false_temp
>> 
>> 
>> ^Cmake[1]: *** [check-js] Interrupt
>> make: *** [distcheck] Interrupt
>> 
>> This is basically taking up my whole night just trying to get past these
>> tests.
>> 
>> 
>> On 3 December 2012 02:02, Noah Slater <nslater@apache.org> wrote:
>> 
>>> (Suffice it to say, this results in unpredictable, and hence useless,
>>> Travis builds.)
>>> 
>>> I am getting this also:
>>> 
>>> ../test/etap/140-attachment-comp.t ......... Failed 9/85 subtests
>>> 
>>> I feel like a broken record at this point, but I do just want to
>>> reitterate: tests that do not work 100% of the time make our testing
>>> infrastructure next to useless. Which may be hyperbole, but false negatives
>>> certainly are not helpful. More to the point, they make my job much harder
>>> than it needs to be. I have literally been trying to get distcheck to pass
>>> for two hours now.
>>> 
>>> The only reason it has taken two hours is that one of the tests fails
>>> each time I run it. I feel like I'm playing the world's most boring and
>>> unfun slot machine. All I need is for this distcheck to pass so I can
>>> continue with my work.
>>> 
>>> For reference, I am doing this on High-Memory Quadruple Extra Large EC2
>>> instance.
>>> 
>>> 
>>> On 3 December 2012 01:53, Noah Slater <nslater@apache.org> wrote:
>>> 
>>>> Hey,
>>>> 
>>>> I am seeing a LOT of this:
>>>> 
>>>> 7624/home/travis/builds/apache/couchdb/apache-couchdb-1.3.0a-0da9587-git/_build/../src/couch_replicator/test/05-replication-many-leaves.t
>>>> .. 22/? Bailout called.  Further testing stopped:  Timeout waiting for
>>>> replication to finish
>>>> 
>>>> Now, this occurs  for me, somewhere towards the end of a distcheck run
>>>> that can take 20 mins. So obviously, each time this happens, it costs me
20
>>>> mins of my time. So that's a little frustrating, heh.
>>>> 
>>>> But it only fails intermittently.
>>>> 
>>>> What can be done?
>>>> 
>>>> Thanks,
>>>> 
>>>> --
>>>> NS
>>>> 
>>> 
>>> 
>>> 
>>> --
>>> NS
>>> 
>> 
>> 
>> 
>> --
>> NS
>> 
> 
> 
> 
> -- 
> NS


Mime
View raw message