You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@couchdb.apache.org by Dave Cottlehuber <da...@muse.net.nz> on 2012/03/03 00:40:56 UTC
Re: [VOTE] Apache CouchDB 1.2.0 -- Windows Binaries
Hey folks,
We're aborting in line with the 1.2.0 source release, so I've moved
the binaries https://people.apache.org/~dch/dist/1.2.0/rc2/
Thanks very much for your testing -- this week has been chaotic with
our move to Vienna so I hope to get some proper couch time again next
week.
Dirkjan, Alex - any progress to share?
A+
Dave
On 26 February 2012 18:46, Andrew Hamilton <ah...@facilityone.com> wrote:
>
> On Feb 25, 2012, at 2:45 PM, Dave Cottlehuber wrote:
>
>> On 24 February 2012 21:18, Andrew Hamilton <ah...@facilityone.com> wrote:
>>> Not sure if this the best way to show my error log, but I am attaching the test suite results from couch.
>>
>> Great - ideal in fact I'd not thought of this myself :-0.
>>
>>> -1
>>> Windows 7 x64
>>> Firefox 11.0
>>> signature OK
>>> md5 & sha OK
>>> No malware
>>> End-user verification OK
>>> Futon tests failed:
>>> attachments
>>> attachment_paths
>>> config
>>> cookie_auth
>>> delayed_commits
>>> design_docs
>>> erlang views
>>> invalid_docids
>>> replicator_db
>>> view_errors
>>
>> Looks like a combo of whitelisting not cleaning up & the same issues
>> already reported prior.
>>
>> Do you want to try same steps as Jeroen (clean up var/lib/couchdb/*
>> and any changes in local.ini, restart and run tests over LAN instead
>> of localhost with local.ini bind_address updated appropriately?
>>
>> This introduces a smidgen of latency which seems to help things.
>
> I ran the test suite again, this time from my home over a VPN to my work computer. I got all tests to pass except for stats, which failed with this message:
>
> Assertion 'triggered, "We managed to force a all_dbs_active error."' failed: We managed to force a all_dbs_active error.
>
> Out of curiosity, I ran this test on localhost, and it passes just fine. Oh, and I forgot to mention earlier I am running the R15B version.
>
>>
>> I am wondering if the issues disappear if POST _restart has a small
>> delay introduced, I'll see if there's a nice way to do this.
>>
>> BTW how fast is this box you are running it on? I'm doing everything
>> on i7 8 core or EC2 large instances. Perhaps this crops up more often
>> on low-end VMs?
>
> My machine is running a Xeon E31220 @ 3.1 GHz and 8GB of RAM. It does seem a lot faster than even some of the servers I have previously tested on.
>
> I'm changing my vote to +1
>
>>
>> A+
>> Dave
>
Re: [VOTE] Apache CouchDB 1.2.0 -- Windows Binaries
Posted by Dave Cottlehuber <da...@muse.net.nz>.
On 4 March 2012 00:49, Noah Slater <ns...@tumbolia.org> wrote:
> I disagree. The old binary builds should be deleted by everyone. There is
> absolutely no reason to keep them around. The source release test procedure
> starts with a command to wipe your existing download and build directory.
>
> On Sat, Mar 3, 2012 at 11:05 PM, Dave Cottlehuber <da...@muse.net.nz> wrote:
>
>> On 3 March 2012 02:05, Noah Slater <ns...@tumbolia.org> wrote:
>> > Dave,
>> >
>> > Please leave the binaries in place, or remove them. I, myself, leave them
>> > in place. I then replace them with for the next round. Please do not use
>> > "rcX" anywhere in the release artefact name. It sends the wrong message.
>> >
>> > Thanks!
>>
>> Fixed by removing them from p.a.o. They are still available in
>> https://www.dropbox.com/s/jeifcxpbtpo78ak/Snapshots/20120223/ for the
>> desperate.
>>
>> Having two binary releases with the same name is even more confusing
>> for people who have downloaded rejected builds. A better approach is
>> needed.
>>
>> Cheers
>> Dave
Wiki updated to reflect nslater's point.
A+
Dave
Re: [VOTE] Apache CouchDB 1.2.0 -- Windows Binaries
Posted by Noah Slater <ns...@tumbolia.org>.
I disagree. The old binary builds should be deleted by everyone. There is
absolutely no reason to keep them around. The source release test procedure
starts with a command to wipe your existing download and build directory.
On Sat, Mar 3, 2012 at 11:05 PM, Dave Cottlehuber <da...@muse.net.nz> wrote:
> On 3 March 2012 02:05, Noah Slater <ns...@tumbolia.org> wrote:
> > Dave,
> >
> > Please leave the binaries in place, or remove them. I, myself, leave them
> > in place. I then replace them with for the next round. Please do not use
> > "rcX" anywhere in the release artefact name. It sends the wrong message.
> >
> > Thanks!
>
> Fixed by removing them from p.a.o. They are still available in
> https://www.dropbox.com/s/jeifcxpbtpo78ak/Snapshots/20120223/ for the
> desperate.
>
> Having two binary releases with the same name is even more confusing
> for people who have downloaded rejected builds. A better approach is
> needed.
>
> Cheers
> Dave
>
Re: [VOTE] Apache CouchDB 1.2.0 -- Windows Binaries
Posted by Dave Cottlehuber <da...@muse.net.nz>.
On 3 March 2012 02:05, Noah Slater <ns...@tumbolia.org> wrote:
> Dave,
>
> Please leave the binaries in place, or remove them. I, myself, leave them
> in place. I then replace them with for the next round. Please do not use
> "rcX" anywhere in the release artefact name. It sends the wrong message.
>
> Thanks!
Fixed by removing them from p.a.o. They are still available in
https://www.dropbox.com/s/jeifcxpbtpo78ak/Snapshots/20120223/ for the
desperate.
Having two binary releases with the same name is even more confusing
for people who have downloaded rejected builds. A better approach is
needed.
Cheers
Dave
Re: [VOTE] Apache CouchDB 1.2.0 -- Windows Binaries
Posted by Noah Slater <ns...@tumbolia.org>.
Dave,
Please leave the binaries in place, or remove them. I, myself, leave them
in place. I then replace them with for the next round. Please do not use
"rcX" anywhere in the release artefact name. It sends the wrong message.
Thanks!
On Fri, Mar 2, 2012 at 11:40 PM, Dave Cottlehuber <da...@muse.net.nz> wrote:
> Hey folks,
>
> We're aborting in line with the 1.2.0 source release, so I've moved
> the binaries https://people.apache.org/~dch/dist/1.2.0/rc2/
>
> Thanks very much for your testing -- this week has been chaotic with
> our move to Vienna so I hope to get some proper couch time again next
> week.
>
> Dirkjan, Alex - any progress to share?
>
> A+
> Dave
>
> On 26 February 2012 18:46, Andrew Hamilton <ah...@facilityone.com>
> wrote:
> >
> > On Feb 25, 2012, at 2:45 PM, Dave Cottlehuber wrote:
> >
> >> On 24 February 2012 21:18, Andrew Hamilton <ah...@facilityone.com>
> wrote:
> >>> Not sure if this the best way to show my error log, but I am attaching
> the test suite results from couch.
> >>
> >> Great - ideal in fact I'd not thought of this myself :-0.
> >>
> >>> -1
> >>> Windows 7 x64
> >>> Firefox 11.0
> >>> signature OK
> >>> md5 & sha OK
> >>> No malware
> >>> End-user verification OK
> >>> Futon tests failed:
> >>> attachments
> >>> attachment_paths
> >>> config
> >>> cookie_auth
> >>> delayed_commits
> >>> design_docs
> >>> erlang views
> >>> invalid_docids
> >>> replicator_db
> >>> view_errors
> >>
> >> Looks like a combo of whitelisting not cleaning up & the same issues
> >> already reported prior.
> >>
> >> Do you want to try same steps as Jeroen (clean up var/lib/couchdb/*
> >> and any changes in local.ini, restart and run tests over LAN instead
> >> of localhost with local.ini bind_address updated appropriately?
> >>
> >> This introduces a smidgen of latency which seems to help things.
> >
> > I ran the test suite again, this time from my home over a VPN to my work
> computer. I got all tests to pass except for stats, which failed with this
> message:
> >
> > Assertion 'triggered, "We managed to force a all_dbs_active error."'
> failed: We managed to force a all_dbs_active error.
> >
> > Out of curiosity, I ran this test on localhost, and it passes just fine.
> Oh, and I forgot to mention earlier I am running the R15B version.
> >
> >>
> >> I am wondering if the issues disappear if POST _restart has a small
> >> delay introduced, I'll see if there's a nice way to do this.
> >>
> >> BTW how fast is this box you are running it on? I'm doing everything
> >> on i7 8 core or EC2 large instances. Perhaps this crops up more often
> >> on low-end VMs?
> >
> > My machine is running a Xeon E31220 @ 3.1 GHz and 8GB of RAM. It does
> seem a lot faster than even some of the servers I have previously tested on.
> >
> > I'm changing my vote to +1
> >
> >>
> >> A+
> >> Dave
> >
>