You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@couchdb.apache.org by Robert Newson <rn...@apache.org> on 2011/05/17 16:17:01 UTC
[VOTE] Apache CouchDB 1.1.0 release
Hello,
I would like call a vote for the Apache CouchDB 1.1.0 release, round 1.
We encourage the whole community to download and test these release artifacts so
that any critical issues can be resolved before the release is made. Everyone is
free to vote on this release, so get stuck in!
We are voting on the following release artifacts:
http://people.apache.org/~rnewson/dist/1.1.0/
These artifacts have been built from the 1.1.0 tag in Subversion:
http://svn.apache.org/repos/asf/couchdb/tags/1.1.0
Happy voting,
--
Robert Newson
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Jan Lehnardt <ja...@apache.org>.
On 17 May 2011, at 17:14, Dirkjan Ochtman wrote:
> On Tue, May 17, 2011 at 16:17, Robert Newson <rn...@apache.org> wrote:
>> We are voting on the following release artifacts:
>>
>> http://people.apache.org/~rnewson/dist/1.1.0/
>
> I tested it on amd64 Gentoo Linux. etap tests passed, browser tests
> mostly passed. I got these failures, but I'm running this on a recent
> Firefox 5.0 alpha:
Hehe, can you try Firefox 3.6? :)
4 or later are currently not supported.
Cheers
Jan
--
>
> oauth
> Exception raised: {"error":"db_not_found","reason":"could not open
> http://localhost:15984/test_suite_db_a/"}
>
> replication
> Exception raised: {"error":"db_not_found","reason":"could not open
> http://localhost:15984/test_suite_db_b/"}
>
> replicator_db
> Exception raised: {}
>
> security_validation
> Exception raised: {"error":"db_not_found","reason":"could not open
> http://localhost:15984/test_suite_db_b/"}
>
> The failures in oauth, replication and security_validation are the
> same as in 1.0.2 in this setup. The replicator_db obviously didn't
> exist in 1.0.2, so not sure about that.
>
> Cheers,
>
> Dirkjan
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Dirkjan Ochtman <di...@ochtman.nl>.
On Tue, May 17, 2011 at 16:17, Robert Newson <rn...@apache.org> wrote:
> We are voting on the following release artifacts:
>
> http://people.apache.org/~rnewson/dist/1.1.0/
I tested it on amd64 Gentoo Linux. etap tests passed, browser tests
mostly passed. I got these failures, but I'm running this on a recent
Firefox 5.0 alpha:
oauth
Exception raised: {"error":"db_not_found","reason":"could not open
http://localhost:15984/test_suite_db_a/"}
replication
Exception raised: {"error":"db_not_found","reason":"could not open
http://localhost:15984/test_suite_db_b/"}
replicator_db
Exception raised: {}
security_validation
Exception raised: {"error":"db_not_found","reason":"could not open
http://localhost:15984/test_suite_db_b/"}
The failures in oauth, replication and security_validation are the
same as in 1.0.2 in this setup. The replicator_db obviously didn't
exist in 1.0.2, so not sure about that.
Cheers,
Dirkjan
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Jan Lehnardt <ja...@apache.org>.
Gabor,
no hurt feelings, there's always a first time :)
Thanks for your enthusiasm. I hope we could channel it into the right directions.
Looking forward to your future contributions :)
Cheers
Jan
--
On 17 May 2011, at 17:25, Gabor Ratky wrote:
> Please fully ignore the email I sent below. As others have pointed out here and on #couchdb, the email has gotten the following things wrong:
> I am not meant to test the release artifacts with the release artifact prepared for voting.
> The release artifacts should be built from source using the tarball provided.
> The purpose of voting is to perform 'make check' and run JS tests in the browser (FF3 suggested in this case)
> The version is correctly stated by `couchdb -V` built from the release artifacts
>
> the minor detail I got right:
> CouchDB 1.1.0 can read 1.0.2 databases
> CouchDB built from the SVN 1.1.0 tag reports 1.2.0a-, opposed to 1.1.0
>
> My only excuse is my excitement of the 1.1.0 release moving forward and rushing to 'help' fellow CouchDB enthusiasts.
>
>
> Let the voting continue (while I leave the stage).
>
> Gabor
> On Tuesday, May 17, 2011 at 4:38 PM, Gabor Ratky wrote:
> Those who develop on Mac OS X and have installed CouchDB using Homebrew [1], should do the following to test CouchDB 1.1.0 with their applications:
>>
>>
>>
>
> redracted to prevent further embarassment
>> On Tuesday, May 17, 2011 at 4:17 PM, Robert Newson wrote:
>>> Hello,
>>>
>>> I would like call a vote for the Apache CouchDB 1.1.0 release, round 1.
>>>
>>> We encourage the whole community to download and test these release artifacts so
>>> that any critical issues can be resolved before the release is made. Everyone is
>>> free to vote on this release, so get stuck in!
>>>
>>> We are voting on the following release artifacts:
>>>
>>> http://people.apache.org/~rnewson/dist/1.1.0/
>>>
>>> These artifacts have been built from the 1.1.0 tag in Subversion:
>>>
>>> http://svn.apache.org/repos/asf/couchdb/tags/1.1.0
>>>
>>> Happy voting,
>>>
>>> --
>>> Robert Newson
>>>
>>
RE: [VOTE] Apache CouchDB 1.1.0 release
Posted by Gabor Ratky <rg...@rgabostyle.com>.
Please fully ignore the email I sent below. As others have pointed out here and on #couchdb, the email has gotten the following things wrong:
I am not meant to test the release artifacts with the release artifact prepared for voting.
The release artifacts should be built from source using the tarball provided.
The purpose of voting is to perform 'make check' and run JS tests in the browser (FF3 suggested in this case)
The version is correctly stated by `couchdb -V` built from the release artifacts
the minor detail I got right:
CouchDB 1.1.0 can read 1.0.2 databases
CouchDB built from the SVN 1.1.0 tag reports 1.2.0a-, opposed to 1.1.0
My only excuse is my excitement of the 1.1.0 release moving forward and rushing to 'help' fellow CouchDB enthusiasts.
Let the voting continue (while I leave the stage).
Gabor
On Tuesday, May 17, 2011 at 4:38 PM, Gabor Ratky wrote:
Those who develop on Mac OS X and have installed CouchDB using Homebrew [1], should do the following to test CouchDB 1.1.0 with their applications:
>
>
>
redracted to prevent further embarassment
> On Tuesday, May 17, 2011 at 4:17 PM, Robert Newson wrote:
> > Hello,
> >
> > I would like call a vote for the Apache CouchDB 1.1.0 release, round 1.
> >
> > We encourage the whole community to download and test these release artifacts so
> > that any critical issues can be resolved before the release is made. Everyone is
> > free to vote on this release, so get stuck in!
> >
> > We are voting on the following release artifacts:
> >
> > http://people.apache.org/~rnewson/dist/1.1.0/
> >
> > These artifacts have been built from the 1.1.0 tag in Subversion:
> >
> > http://svn.apache.org/repos/asf/couchdb/tags/1.1.0
> >
> > Happy voting,
> >
> > --
> > Robert Newson
> >
>
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Robert Newson <ro...@gmail.com>.
also this;
PS It seems the version number has not been set correctly yet,
`couchdb -V` returns 'couchdb - Apache CouchDB 1.2.0a-' for me
appears not to be the case. The build artifact correctly reports 1.1.0
for me and others.
B.
On 17 May 2011 15:44, Noah Slater <ns...@apache.org> wrote:
> On 17 May 2011, at 15:38, Gabor Ratky wrote:
>
>> Those who develop on Mac OS X and have installed CouchDB using Homebrew [1], should do the following to test CouchDB 1.1.0 with their applications:
>
> If you're using CouchDB on your development machine, I would strongly recommend AGAINST doing this. This release artefact is only to be used for voting on at the moment. It is not intended to be used by anyone properly, or to be used on any real data.
>
>
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
On 17 May 2011, at 15:38, Gabor Ratky wrote:
> Those who develop on Mac OS X and have installed CouchDB using Homebrew [1], should do the following to test CouchDB 1.1.0 with their applications:
If you're using CouchDB on your development machine, I would strongly recommend AGAINST doing this. This release artefact is only to be used for voting on at the moment. It is not intended to be used by anyone properly, or to be used on any real data.
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
On 17 May 2011, at 15:38, Gabor Ratky wrote:
> Those who develop on Mac OS X and have installed CouchDB using Homebrew [1], should do the following to test CouchDB 1.1.0 with their applications:
Also, please do not test the release artefact in this way. If you would like to vote on the release, please build from source directly, and by hand. That is absolutely crucial for the process to work properly.
(Sorry for the negative feedback Gabor, I know you were trying to help. I just don't think that this particular approach is suitable for what we're trying to do here...)
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Robert Newson <ro...@gmail.com>.
And also 1.1.0 can read 1.0.2 databases though, as Noah noted, you
shouldn't be trying this artifact on data you care about (a copy is
fine, obviously).
B.
On 17 May 2011 15:47, Jan Lehnardt <ja...@apache.org> wrote:
>
> On 17 May 2011, at 16:38, Gabor Ratky wrote:
>
>> PS It seems the version number has not been set correctly yet, `couchdb -V` returns 'couchdb - Apache CouchDB 1.2.0a-' for me
>
> It seems you may be testing a binary that is derived from trunk, not the branch or posted tarball.
>
> Cheers
> Jan
> --
>
>
>>
>> [1] http://github.com/mxcl/homebrew (if you are not using Homebrew yet, you should seriously consider!)
>> On Tuesday, May 17, 2011 at 4:17 PM, Robert Newson wrote:
>>> Hello,
>>>
>>> I would like call a vote for the Apache CouchDB 1.1.0 release, round 1.
>>>
>>> We encourage the whole community to download and test these release artifacts so
>>> that any critical issues can be resolved before the release is made. Everyone is
>>> free to vote on this release, so get stuck in!
>>>
>>> We are voting on the following release artifacts:
>>>
>>> http://people.apache.org/~rnewson/dist/1.1.0/
>>>
>>> These artifacts have been built from the 1.1.0 tag in Subversion:
>>>
>>> http://svn.apache.org/repos/asf/couchdb/tags/1.1.0
>>>
>>> Happy voting,
>>>
>>> --
>>> Robert Newson
>>>
>
>
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Jan Lehnardt <ja...@apache.org>.
On 17 May 2011, at 16:38, Gabor Ratky wrote:
> PS It seems the version number has not been set correctly yet, `couchdb -V` returns 'couchdb - Apache CouchDB 1.2.0a-' for me
It seems you may be testing a binary that is derived from trunk, not the branch or posted tarball.
Cheers
Jan
--
>
> [1] http://github.com/mxcl/homebrew (if you are not using Homebrew yet, you should seriously consider!)
> On Tuesday, May 17, 2011 at 4:17 PM, Robert Newson wrote:
>> Hello,
>>
>> I would like call a vote for the Apache CouchDB 1.1.0 release, round 1.
>>
>> We encourage the whole community to download and test these release artifacts so
>> that any critical issues can be resolved before the release is made. Everyone is
>> free to vote on this release, so get stuck in!
>>
>> We are voting on the following release artifacts:
>>
>> http://people.apache.org/~rnewson/dist/1.1.0/
>>
>> These artifacts have been built from the 1.1.0 tag in Subversion:
>>
>> http://svn.apache.org/repos/asf/couchdb/tags/1.1.0
>>
>> Happy voting,
>>
>> --
>> Robert Newson
>>
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
On 17 May 2011, at 15:38, Gabor Ratky wrote:
> PS It seems the version number has not been set correctly yet, `couchdb -V` returns 'couchdb - Apache CouchDB 1.2.0a-' for me
We've not been able to reproduce this.
Could you provide some more information?
What does curl localhost:5984 produce?
RE: [VOTE] Apache CouchDB 1.1.0 release
Posted by Gabor Ratky <rg...@rgabostyle.com>.
Those who develop on Mac OS X and have installed CouchDB using Homebrew [1], should do the following to test CouchDB 1.1.0 with their applications:
`brew update` # Make sure you have the latest formulae
`brew edit couchdb` # Edit the CouchDB formula using $EDITOR
and replace the following line:
url 'https://github.com/apache/couchdb/tarball/1.0.2'
with:
url 'http://people.apache.org/~rnewson/dist/1.1.0/apache-couchdb-1.1.0.tar.gz' # can also use 'http://svn.apache.org/repos/asf/couchdb/tags/1.1.0' if you'd rather do an SVN checkout of the 1.1.0 tag
Finally,
`brew remove couchdb` # Remove the installed stable CouchDB 1.0.2
`brew install couchdb` # Build and install CouchDB 1.1.0 from Bob's 1.1.0 tarball
All databases will be intact and immediately available with CouchDB 1.1.0. Subsequently removing CouchDB 1.1.0 and undoing the changes made to the formula will make it easy to go back to 1.0.2. Later, when the official release happens, the canonical formula will be updated by someone and signed off @mxcl or @adamv (core maintainers of Homebrew).
The only thing that might affect your development environment is the new database format the CouchDB 1.1.0 migrates to during a database compaction. Those database files will not be understood by 1.0.2 and will have to be recreated or replicated from a backup database (possibly between different versions of CouchDB).
Let me know if you have questions or something is not clear, I'm more than happy to help with Homebrew / CouchDB build/install issues.
Gabor
PS It seems the version number has not been set correctly yet, `couchdb -V` returns 'couchdb - Apache CouchDB 1.2.0a-' for me
[1] http://github.com/mxcl/homebrew (if you are not using Homebrew yet, you should seriously consider!)
On Tuesday, May 17, 2011 at 4:17 PM, Robert Newson wrote:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.1.0 release, round 1.
>
> We encourage the whole community to download and test these release artifacts so
> that any critical issues can be resolved before the release is made. Everyone is
> free to vote on this release, so get stuck in!
>
> We are voting on the following release artifacts:
>
> http://people.apache.org/~rnewson/dist/1.1.0/
>
> These artifacts have been built from the 1.1.0 tag in Subversion:
>
> http://svn.apache.org/repos/asf/couchdb/tags/1.1.0
>
> Happy voting,
>
> --
> Robert Newson
>
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Jan Lehnardt <ja...@apache.org>.
On 17 May 2011, at 16:17, Robert Newson wrote:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.1.0 release, round 1.
>
> We encourage the whole community to download and test these release artifacts so
> that any critical issues can be resolved before the release is made. Everyone is
> free to vote on this release, so get stuck in!
>
> We are voting on the following release artifacts:
>
> http://people.apache.org/~rnewson/dist/1.1.0/
>
> These artifacts have been built from the 1.1.0 tag in Subversion:
>
> http://svn.apache.org/repos/asf/couchdb/tags/1.1.0
>
> Happy voting,
+1
Tested on Mac OS X 10.6.7, R14B01. make check and the test suite in Firefox 3.6 all check out.
Tested on Ubuntu 10.04, R13B03. make check and the test suite in Firefox 3.6 all check out.
Thanks Bob!
Cheers
Jan
--
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
On 17 May 2011, at 18:21, Paul Davis wrote:
> Can you reproduce it? Which test was running when it failed? If you
> can reproduce it, can you attach gdb to your couchdb process to get a
> stack trace of when it segfaults?
Yep, it's the design_docs tests. Segfaults every time.
From the log:
Apache CouchDB 1.1.0 (LogLevel=info) is starting.
Apache CouchDB has started. Time to relax.
[info] [<0.1248.0>] Apache CouchDB has started on http://127.0.0.1:5984/
[info] [<0.1331.0>] 127.0.0.1 - - 'GET' / 200
[info] [<0.1331.0>] 127.0.0.1 - - 'GET' / 200
[info] [<0.1367.0>] OS Process #Port<0.11151> Log :: function raised exception (new ReferenceError("makebigstring is not defined", "", 2)) with doc._id 1
[info] [<0.1358.0>] checkpointing view update at seq 4 for test_suite_db _design/test
[info] [<0.1358.0>] checkpointing view update at seq 289 for test_suite_db _design/test
[info] [<0.1358.0>] checkpointing view update at seq 716 for test_suite_db _design/test
[info] [<0.1358.0>] checkpointing view update at seq 1003 for test_suite_db _design/test
[info] [<0.1331.0>] 127.0.0.1 - - 'GET' /test_suite_db/_design/test/_view/commonjs?limit=1 200
[info] [<0.1331.0>] 127.0.0.1 - - 'GET' /test_suite_db/_all_docs?startkey=%22_design%22&endkey=%22_design0%22 200
[info] [<0.1331.0>] 127.0.0.1 - - 'GET' /test_suite_db/_design/test/_view/all_docs_twice 200
[info] [<0.1331.0>] 127.0.0.1 - - 'GET' /test_suite_db/_design/test/_view/no_docs 200
[info] [<0.1331.0>] 127.0.0.1 - - 'GET' /test_suite_db/_design/test/_view/single_doc 200
[info] [<0.1331.0>] 127.0.0.1 - - 'POST' /test_suite_db/_ensure_full_commit 201
[info] [<0.1331.0>] 127.0.0.1 - - 'POST' /_restart 200
[info] [<0.1314.0>] Stopping all ongoing replications because the replicator database was deleted or changed
Apache CouchDB 1.1.0 (LogLevel=info) is starting.
Segmentation fault
My threats to debug it with GDB cause the problem to go away.
Damn heisenbugs.
I'll keep trying.
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Robert Newson <ro...@gmail.com>.
All,
Firstly, thank you to everyone who took the time to vote on this artifact.
Unfortunately I'm voting -1 and aborting round one of this release.
We've uncovered a bug serious enough that we would like a fix for it
to ship with 1.1.0 (COUCHDB-885). The behavior at fault is not
verified by any of our automated tests, which is why no one in this
thread has seen it. This bug appears to have been present in all
releases since 0.10.2 but is not present on trunk (0.11.x has a
slightly different replicator to 0.10.x, and trunk has a whole new
one).
Round 2 will begin once we have fixed 885 and any related tickets.
Thanks again!
B.
On 17 May 2011 18:27, Paul Davis <pa...@gmail.com> wrote:
> On Tue, May 17, 2011 at 1:21 PM, Paul Davis <pa...@gmail.com> wrote:
>> On Tue, May 17, 2011 at 1:15 PM, Noah Slater <ns...@apache.org> wrote:
>>> I'm getting the following unit test failures:
>>>
>>> http://friendpaste.com/ZMbw7g2OxETvcAzEDDcGv
>>>
>>
>> This is a known race condition. The problem is the test, not what's
>> being tested.
>>
>>> I tried running CouchDB and got a segfault.
>>>
>>
>> Can you reproduce it? Which test was running when it failed? If you
>> can reproduce it, can you attach gdb to your couchdb process to get a
>> stack trace of when it segfaults?
>>
>>> I'm on OS X, latest version, yada yada.
>>>
>>> Let me know if you need any more information.
>>>
>>
>
> I'm getting an error in Futon in Chrome in the replicator test:
>
> Assertion failed: dbB.open("30") !== null
>
> Also I had it fail once so that an admin was left configured which
> cause every other test to fail.
>
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Paul Davis <pa...@gmail.com>.
On Tue, May 17, 2011 at 1:21 PM, Paul Davis <pa...@gmail.com> wrote:
> On Tue, May 17, 2011 at 1:15 PM, Noah Slater <ns...@apache.org> wrote:
>> I'm getting the following unit test failures:
>>
>> http://friendpaste.com/ZMbw7g2OxETvcAzEDDcGv
>>
>
> This is a known race condition. The problem is the test, not what's
> being tested.
>
>> I tried running CouchDB and got a segfault.
>>
>
> Can you reproduce it? Which test was running when it failed? If you
> can reproduce it, can you attach gdb to your couchdb process to get a
> stack trace of when it segfaults?
>
>> I'm on OS X, latest version, yada yada.
>>
>> Let me know if you need any more information.
>>
>
I'm getting an error in Futon in Chrome in the replicator test:
Assertion failed: dbB.open("30") !== null
Also I had it fail once so that an admin was left configured which
cause every other test to fail.
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
On 17 May 2011, at 23:35, Jan Lehnardt wrote:
>> This is the most surreal email I've read for a long time.
>
> I work with this guy, it's like that every day.
Just for the record, the genesis of this meme was on IRC today.
18:54 <+rnewson> davisp: I sent a manila envelope full of segfaults to nslater the other day, looks like they arrived.
18:55 <+davisp> rnewson: Had any kernel panics lately?
18:55 <+rnewson> davisp: nope.
18:55 <+rnewson> davisp: Oh, expect a parcel soon.
18:55 <+davisp> heh
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Jan Lehnardt <ja...@apache.org>.
On 18 May 2011, at 00:34, Noah Slater wrote:
>
> On 17 May 2011, at 23:30, Chris Anderson wrote:
>
>> +1 on the manilla envelope. I'm super happy that we have momentum on
>> the release process. And it looks like some of the fixes we are
>> finding are useful.
>>
>> It may be frustrating, but this is exactly what a release process is
>> for -- finding those manila envelopes, and fixing them.
>
>
> This is the most surreal email I've read for a long time.
I work with this guy, it's like that every day.
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
On 17 May 2011, at 23:30, Chris Anderson wrote:
> +1 on the manilla envelope. I'm super happy that we have momentum on
> the release process. And it looks like some of the fixes we are
> finding are useful.
>
> It may be frustrating, but this is exactly what a release process is
> for -- finding those manila envelopes, and fixing them.
This is the most surreal email I've read for a long time.
Hehe.
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Chris Anderson <jc...@apache.org>.
+1 on the manilla envelope. I'm super happy that we have momentum on
the release process. And it looks like some of the fixes we are
finding are useful.
It may be frustrating, but this is exactly what a release process is
for -- finding those manila envelopes, and fixing them.
Chris
On Tue, May 17, 2011 at 12:27 PM, Noah Slater <ns...@apache.org> wrote:
>
> On 17 May 2011, at 20:24, Paul Davis wrote:
>
>> Let me know when your manilla envelope of segfaults is empty.
>
> I have stopped testing for now, as I do not know how to proceed.
>
> I'm sending you a manilla envelope of discontentment.
>
>
--
Chris Anderson
http://jchrisa.net
http://couchbase.com
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
On 17 May 2011, at 20:24, Paul Davis wrote:
> Let me know when your manilla envelope of segfaults is empty.
I have stopped testing for now, as I do not know how to proceed.
I'm sending you a manilla envelope of discontentment.
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Paul Davis <pa...@gmail.com>.
On Tue, May 17, 2011 at 3:18 PM, Noah Slater <ns...@apache.org> wrote:
>
> On 17 May 2011, at 19:49, Paul Davis wrote:
>
>> Maybe your fuzzy logic board needs replaced?
>
> New MBA is new.
>
> I installed by doing "brew install couchdb" and then installing CouchDB from source from the tarball. Jan recomended this as a way of getting the dependancies installed with minimum fuss. Open to suggestions on how to test around this. Going to -1 the vote unless it passes for me. (Or someone convinces me my brand new MBA is some how a unique and special snowflake that randomly causes segfaults with stable software.)
Let me know when your manilla envelope of segfaults is empty.
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
On 17 May 2011, at 19:49, Paul Davis wrote:
> Maybe your fuzzy logic board needs replaced?
New MBA is new.
I installed by doing "brew install couchdb" and then installing CouchDB from source from the tarball. Jan recomended this as a way of getting the dependancies installed with minimum fuss. Open to suggestions on how to test around this. Going to -1 the vote unless it passes for me. (Or someone convinces me my brand new MBA is some how a unique and special snowflake that randomly causes segfaults with stable software.)
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Paul Davis <pa...@gmail.com>.
On Tue, May 17, 2011 at 2:02 PM, Noah Slater <ns...@apache.org> wrote:
>
> On 17 May 2011, at 18:21, Paul Davis wrote:
>
>> Can you reproduce it? Which test was running when it failed? If you
>> can reproduce it, can you attach gdb to your couchdb process to get a
>> stack trace of when it segfaults?
>
> Okay, now we have:
>
> Program received signal EXC_BAD_ACCESS, Could not access memory.
> Reason: KERN_INVALID_ADDRESS at address: 0x00000000107c3bc0
> [Switching to process 99177 thread 0x2a03]
> 0x00000000107c3bc0 in ?? ()
> (gdb)
> Continuing.
>
> Program received signal EXC_BAD_ACCESS, Could not access memory.
> Reason: KERN_INVALID_ADDRESS at address: 0x00000000107c3bc0
> 0x00000000107c3bc0 in ?? ()
> (gdb) bt
> #0 0x00000000107c3bc0 in ?? ()
> Cannot access memory at address 0x107c3bc0
> #1 0x00007fff86abe5b1 in EVP_DigestInit_ex ()
> #2 0x00007fff86a8c89d in ssleay_rand_bytes ()
> #3 0x00007fff86a8bf2e in ssleay_rand_pseudo_bytes ()
> #4 0x000000001074307a in rand_bytes_1 ()
> #5 0x00000000100fdc07 in process_main ()
> #6 0x000000001006ff5e in sched_thread_func ()
> #7 0x000000001018e8db in thr_wrapper ()
> #8 0x00007fff8909f4f6 in _pthread_start ()
> #9 0x00007fff8909f3a9 in thread_start ()
> (gdb)
>
>
>
That's down in OpenSSL in what appears to be a call to
`crypto:rand_bytes(N)`. In other words, something goofy is going on
cause that bit of code is pretty core to couchdb and the various
tests, so if it were widespread we would've heard reports.
Maybe your fuzzy logic board needs replaced?
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
On 17 May 2011, at 18:21, Paul Davis wrote:
> Can you reproduce it? Which test was running when it failed? If you
> can reproduce it, can you attach gdb to your couchdb process to get a
> stack trace of when it segfaults?
Okay, now we have:
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0x00000000107c3bc0
[Switching to process 99177 thread 0x2a03]
0x00000000107c3bc0 in ?? ()
(gdb)
Continuing.
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0x00000000107c3bc0
0x00000000107c3bc0 in ?? ()
(gdb) bt
#0 0x00000000107c3bc0 in ?? ()
Cannot access memory at address 0x107c3bc0
#1 0x00007fff86abe5b1 in EVP_DigestInit_ex ()
#2 0x00007fff86a8c89d in ssleay_rand_bytes ()
#3 0x00007fff86a8bf2e in ssleay_rand_pseudo_bytes ()
#4 0x000000001074307a in rand_bytes_1 ()
#5 0x00000000100fdc07 in process_main ()
#6 0x000000001006ff5e in sched_thread_func ()
#7 0x000000001018e8db in thr_wrapper ()
#8 0x00007fff8909f4f6 in _pthread_start ()
#9 0x00007fff8909f3a9 in thread_start ()
(gdb)
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Paul Davis <pa...@gmail.com>.
On Tue, May 17, 2011 at 1:15 PM, Noah Slater <ns...@apache.org> wrote:
> I'm getting the following unit test failures:
>
> http://friendpaste.com/ZMbw7g2OxETvcAzEDDcGv
>
This is a known race condition. The problem is the test, not what's
being tested.
> I tried running CouchDB and got a segfault.
>
Can you reproduce it? Which test was running when it failed? If you
can reproduce it, can you attach gdb to your couchdb process to get a
stack trace of when it segfaults?
> I'm on OS X, latest version, yada yada.
>
> Let me know if you need any more information.
>
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
I'm getting the following unit test failures:
http://friendpaste.com/ZMbw7g2OxETvcAzEDDcGv
I tried running CouchDB and got a segfault.
I'm on OS X, latest version, yada yada.
Let me know if you need any more information.
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Noah Slater <ns...@apache.org>.
I have created a test procedure:
http://wiki.apache.org/couchdb/Test_procedure
This should help new users figure out how they can help with the vote.
Please let me know your thoughts on it.
Probably made a few mistakes, or missed some things out.
Re: [VOTE] Apache CouchDB 1.1.0 release
Posted by Matt Goodall <ma...@gmail.com>.
On 17 May 2011 15:17, Robert Newson <rn...@apache.org> wrote:
> Hello,
>
> I would like call a vote for the Apache CouchDB 1.1.0 release, round 1.
>
> We encourage the whole community to download and test these release
> artifacts so
> that any critical issues can be resolved before the release is made.
> Everyone is
> free to vote on this release, so get stuck in!
>
> We are voting on the following release artifacts:
>
> http://people.apache.org/~rnewson/dist/1.1.0/
>
> These artifacts have been built from the 1.1.0 tag in Subversion:
>
> http://svn.apache.org/repos/asf/couchdb/tags/1.1.0
>
> Happy voting,
>
+1
Ubuntu 11.04, built against erlang 1:13.b.3-dfsg-2ubuntu3 and xulrunner
1.9.2.
* make check all ok.
* Test suite failed on 'changes' test in Firefox 4.0.1 on full run but
passed when run on its own.
* Test suite passed in Chromium 12.0.742.30.
- Matt