[typo] Frustrating Dreamhost/Typo experience

[Scott Hughes (List)]

On Aug 25, 2005, at 6:25 PM, Scott Hughes (list) wrote:

> Scott Laird wrote:
>
>
>> On Aug 25, 2005, at 1:11 PM, Scott Hughes (List) wrote:
>>
>>
>>> After a while of waiting for it to "fix itself", I end up doing  
>>> a  "killall -9 dispatch.fcgi;  rake sweep_cache" and wait about  
>>> 15-20  minutes for it to start working again.  I don't know if  
>>> that's the  best procedure for reseting it or not...  And I'm  
>>> really curious as  to why it takes 15-20 minutes to start working  
>>> again.
>>>
>>
>>
>> *15 minutes*.  That's really slow--my home system (Athlon 700, so  
>> not  very fast) is ready to start serving pages within a couple  
>> seconds.   I'd start by figuring out what's happening here.  Try  
>> running  dispatch.fcgi by hand and see if it spits out any errors  
>> or anything  else weird happens.  I don't use dreamhost, so I  
>> don't have any  concrete suggestions, but this seems really strange.
>>
>>
> Scott,
>
> Thanks for your response.  Oddly enough, I saw a ton of errors when  
> I ran "dispatch.fcgi" right after reading your email.  When I  
> switched terminals to compose this response, I couldn't get it to  
> display an error.  I remember that the error message (which was  
> repeated a lot) had something to do with a string conversion on a  
> Nil type, and the bottom of the stack referenced line 24 of  
> dispatch.fcgi.
>

So, as it turns out, my error/non-error evaluation was wrong.  When I  
said (above) that I couldn't get it to display an error, I was fooled  
by seeing the full html of my default.rhtml file.  Had I read that  
html, I would see that it displayed an Application Error (versus when  
I saw a ton of errors, the process hang, and I saw no html in the  
output).  Above the application error I get the following:

<p>undefined method `downcase' for nil:NilClass</p>

Which is the same message I saw repeated numerous times when it was  
locking up on me.

Scott