incubator-couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Keith Gable <zi...@ignition-project.com>
Subject Re: Binary logs?
Date Thu, 27 Oct 2011 19:38:11 GMT
Honestly, I'd rm the log and go on my way. Am I horrible?
On Oct 27, 2011 2:34 PM, "CGS" <cgsmcmlxxv@gmail.com> wrote:

> I don't know. It may be a core dump from Erlang or other things which went
> wrong when you started CouchDB for the first time. It is hard to say if the
> problem cannot be reproduced.
>
>
>
> On 10/27/2011 07:04 PM, Travis Paul wrote:
>
>> Yeah vim shows a screen full of ^@^@^@^@ characters too, deleting them
>> lets
>> me view it as a test file again, but I wonder how it they got there?
>> My other couch logs have those characters as well (i.e couch.log.1)
>>
>> On Thu, Oct 27, 2011 at 12:57 PM, CGS<cgsmcmlxxv@gmail.com>  wrote:
>>
>>  Try to open your file with vim and delete manually those chars at the top
>>> of your file (somehow you accumulated garbage at the top of your file).
>>> Save
>>> it with another name if you want to keep the original (if you want next
>>> time
>>> not to repeat the operation, save it over the old file) and try grep
>>> again.
>>>
>>>
>>>
>>>
>>> On 10/27/2011 06:51 PM, Travis Paul wrote:
>>>
>>>  It is likely an error in my code somewhere, I'm not too concerned about
>>>> those errors as that app is nothing serious and needs to be taken down
>>>> anyways.
>>>>
>>>> As for opening with less I see a bunch of "^@^@^@^@" characters filling
>>>> the
>>>> screen (maybe this means something in less?) and if I scroll down I can
>>>> see
>>>> normal log messages.
>>>>
>>>> I opened the file with hexedit and to be honest, had no idea what I was
>>>> looking for...
>>>>
>>>> Using *strings* or *grep -a* works for me so, I supposed it's not a
>>>> major
>>>>
>>>> issue, I can share the log file if you are interested.
>>>>
>>>> Thanks for your help
>>>>
>>>>
>>>> On Thu, Oct 27, 2011 at 12:33 PM, Robert Newson<rnewson@apache.org>
>>>>  wrote:
>>>>
>>>>  Those are all normal ASCII values, though. You might need to open the
>>>>
>>>>> file in a hex editor to examine it safely. Grep is likely scanning for
>>>>> bytes over 127 at the beginning of the file.
>>>>>
>>>>> As for the error, it looks like couchjs is crashing in your list
>>>>> function, so perhaps you have a syntax error in your code?
>>>>>
>>>>> B.
>>>>>
>>>>> On 27 October 2011 17:30, Travis Paul<Tr@vispaul.me>   wrote:
>>>>>
>>>>>  I do see some errors that are probably making grep think the file is
>>>>>>
>>>>>>  binary
>>>>>
>>>>>  and I pasted them here http://pastebin.com/rA9kHSNF if they are at
>>>>>> all
>>>>>>
>>>>>>  of
>>>>>
>>>>>  interest. I am unsure of their meaning.
>>>>>>
>>>>>> I encountered a 'bizarre' error that I want to catch, which occurred
>>>>>> when
>>>>>>
>>>>>>  I
>>>>>
>>>>>  accessed the _rewrite url  to a view and Couch returned a lot of plain
>>>>>>
>>>>>>  text
>>>>>
>>>>>  that looked like a stack trace, which a page refresh corrected and I
>>>>>> have
>>>>>> not been able to recreate. I don't recall the date when I encountered
>>>>>>
>>>>>>  that
>>>>>
>>>>>  error or what the error message looked like, so I'm not sure what to
>>>>>> look
>>>>>> for in old logs, but I never want a user to see it so I'm trying
to
>>>>>> catch
>>>>>>
>>>>>>  it
>>>>>
>>>>>  to see if it was something that I did or possibly a bug.
>>>>>>
>>>>>>
>>>>>> On Thu, Oct 27, 2011 at 12:06 PM, Robert Newson<rnewson@apache.org>
>>>>>>
>>>>>>  wrote:
>>>>>
>>>>>  Grep has a heuristic to detect text vs. binary by examining the first
>>>>>>
>>>>>>> part of the file. I'm curious to know what you have in, say,
the
>>>>>>> first
>>>>>>> 100 bytes.
>>>>>>>
>>>>>>> couch.log is a text file, my guess is you have some strange
>>>>>>> characters
>>>>>>> in your log that is fooling grep.
>>>>>>>
>>>>>>> B.
>>>>>>>
>>>>>>> On 27 October 2011 16:59, Travis Paul<Tr@vispaul.me>  
wrote:
>>>>>>>
>>>>>>>  If I try grep without the -a flag such as:  *grep "Thu, 27 Oct
2011"
>>>>>>>> couch.log*
>>>>>>>> I get: *Binary file couch.log matches*
>>>>>>>>
>>>>>>>> tail, and head commands work as usual, but if I want to open
the
>>>>>>>> file
>>>>>>>>
>>>>>>>>  in
>>>>>>>
>>>>>> gedit I have to do something like: *strings couch.log>
>>>>>>
>>>>>>> couch.log.text*
>>>>>>>
>>>>>> less command complains as well: *"couch.log" may be a binary file.
>>>>>>
>>>>>>>  See
>>>>>>>
>>>>>> it
>>>>>>
>>>>>>> anyway?*
>>>>>>>>
>>>>>>>> On Thu, Oct 27, 2011 at 11:51 AM, Robert Newson<rnewson@apache.org>
>>>>>>>>
>>>>>>>>  wrote:
>>>>>>>
>>>>>>>  couch.log is a text file with lines like;
>>>>>>>>
>>>>>>>>> [Thu, 27 Oct 2011 10:30:57 GMT] [info] [<0.15397.1>]
127.0.0.1 - -
>>>>>>>>>
>>>>>>>>>  GET
>>>>>>>>
>>>>>>>  /db1/doc1 200
>>>>>>
>>>>>>> What are you seeing in there?
>>>>>>>>>
>>>>>>>>> B.
>>>>>>>>>
>>>>>>>>> On 27 October 2011 16:36, Travis Paul<Tr@vispaul.me>
  wrote:
>>>>>>>>>
>>>>>>>>>  Any idea why the couchdb logs in /var/log/couchdb are
binary
>>>>>>>>>> files?
>>>>>>>>>>
>>>>>>>>>>  I
>>>>>>>>>
>>>>>>>>  have
>>>>>>
>>>>>>> to use *strings* or *grep -a* to do anything with them... just
>>>>>>>>>>
>>>>>>>>>>  seemed
>>>>>>>>>
>>>>>>>>  unusual.
>>>>>>
>>>>>>> Also, does anyone have a custom logwatch service for couchdb
they
>>>>>>>>>>
>>>>>>>>>>  would
>>>>>>>>>
>>>>>>>> like
>>>>>>>>
>>>>>>>>> to share? I'm going to be making one tomorrow if I can't
find an
>>>>>>>>>>
>>>>>>>>>>  existing
>>>>>>>>>
>>>>>>>> service.
>>>>>>>>
>>>>>>>>> Thanks!
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message