incubator-couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Newson <rnew...@apache.org>
Subject Re: Upgrade from 1.1.0 to 1.2.0
Date Thu, 21 Jun 2012 20:08:03 GMT

The extra debug information isn't helpful. What we actually need is the rest of the "Reason
for termination".

I suspect the returned binary is shorter than the number of bytes we expected to read but
I can't confirm that from a partial log. If that turns out to be the case then it's one of
two things. 1) the size of the iolist recorded by couchdb is wrong (too large) 2) the file
has been truncated somehow.

B.

On 21 Jun 2012, at 20:56, Ronnie F. Moller, Jr. wrote:

>> -----Original Message-----
>> From: Dave Cottlehuber [mailto:dave@muse.net.nz]
>> Sent: Wednesday, June 20, 2012 4:50 PM
>> To: user@couchdb.apache.org
>> Subject: Re: Upgrade from 1.1.0 to 1.2.0
>> 
>> On 20 June 2012 23:31, Ronnie F. Moller, Jr.
>> <ronnie@redbaronconsulting.com> wrote:
>>>> -----Original Message-----
>>>> From: Dave Cottlehuber [mailto:dave@muse.net.nz]
>>>> Sent: Wednesday, June 20, 2012 4:12 PM
>>>> To: user@couchdb.apache.org
>>>> Subject: Re: Upgrade from 1.1.0 to 1.2.0
>>>> 
>>>> On 20 June 2012 22:33, Ronnie F. Moller, Jr.
>>>> <ronnie@redbaronconsulting.com> wrote:
>>>>> 
>>>>>> 
>>>>>> I am moving servers and in the process upgrading to 1.2.0.
>>>> 
>>>> I.e. on new server you have 1.2.0 installed, and the old one has
> 1.0.1
>>> ?
>>> 
>>> Yes. Not touching the original old server.
>>>> 
>>>>>> My plan is to move the couchdb files manually. The old server
>>> definition
>>>> has a util driver of couch-1.0.1, so I updated this to 1.2.0.
>>>> 
>>>> Um you mean your edited default.ini? Or this was found in your
>>> local.ini file?
>>>> You shouldn't need to add this to the new server's local.ini
>>> 
>>> The old server had this in the default.ini, and the default.ini is
> what
>>> was modified to point to the new data directory, instead of
> overriding
>>> the local.ini. I guess I could leave the default.ini stock, and do
> just
>>> modify the local.ini where needed. This does produce the same
>>> end-results, as long as I do not include the util-driver line.
>>> 
>>>>>> Now when I run the web interface I get a growing log file and
> the
>>> erlang
>>>> debug window shows lots of numbers just streaming.
>>>>>> 
>>>>>> Is this an upgrade process, or did I make a bad assumption on
> the
>>>> process?
>>>> 
>>>> does curling / work?
>>>> 
>>>> how are you starting couchdb? any immediate errors on startup
>>>> 
>>> 
>>> I am starting couchdb using the "start couch" batch file so that the
>>> debug window for erl is shown. I can browse to the utils, and
> everything
>>> appears normal until I attempt to login as a user. Then it outputs a
>>> constant scrolling of numbers separated by spaces. At first I
> thought
>>> this was an update, until I realized they were not incrementing, but
>>> appear random.
>> 
>> Hi Ronnie,
>> 
>> You're making it *really* hard for us to help here. Mentioning that
> it's
>> running Windows is significant (even I assumed otherwise).
>> 
>> What OS are you running & which build of CouchDB?
>> 
>> Did you verify the installation before moving your DB files in?
>> 
>> -> fresh install of couchdb to an empty directory
>> -> launch any browser except IE
>> -> open futon
>> -> right-hand side is a "verify install" button
>> 
>> Does that work?
>> 
>> If not, restart couchdb & don't run futon, just provide output using
>> 
>>    curl -v http://localhost:5984/
>> 
>> And please copy & paste the output from the werl window somewhere
>> (not in email) please, e.g. friendpaste?
>> 
>> Unfortunately I'm off for the evening here, I'll look tomorrow for
> your
>> update.
>> 
>> A+
>> Dave
>> 
>> Hi Ronnie,
>> 
>> Aaah yes SBS2011 I should not have forgotten so soon :-).
>> 
>> Um can I see the full stack log? In particular a line number will help.
> 
> I set the log level to debug. Not sure if this is what you needed.
> 
> https://friendpaste.com/rGwSwfqwMuEy3tBc8FnY8
> 
>> I'm guessing there's some sort of IO error here, pread_iolist.
>> Is this a fresh install now? If so can you run a chkdsk?
> 
> Fresh install of OS. Ran a chkdsk, and a full reboot with physical scan.
> No errors
> 
> -Ronnie


Mime
View raw message