MogileFS Issues (no_devices No devices found to store file)

[dormando]

I ... missed that conversation? Wasn't on the mailing list, at least. 
I'm also running a few revisions behind, so I would've missed it coming 
down the wire too.

-Dormando

Brad Fitzpatrick wrote:
> Didn't we remove that in a recent version, because it was biting so many
> people?  I thought we now just compared the server's Date: header (in the
> HTTP response) to the Last-Modified: header, and made sure the file was
> *relatively* new enough, compared to itself, not compared to the
> tracker....
> 
> Or did we just talk about that?
> 
> 
> On Thu, 15 Mar 2007, dormando wrote:
> 
>> I've also seen this if the clocks are off between the database and
>> mogile tracker. You can check the host table to see the "age" of each node.
>>
>> -Dormando
>>
>> Travis Boucher wrote:
>>> Nope.
>>>        host device            size(G)    used(G)    free(G)   use%
>>>        ---- --------------- ---------- ---------- ---------- ------
>>>        [ 1] dev1               16.958      0.031     16.927   0.18%
>>> .. 84 times
>>>
>>> (A bunch of small fast drives)
>>>
>>> Brad Fitzpatrick wrote:
>>>> Out of disk space?  Under 100 MB?
>>>>
>>>>
>>>> On Thu, 15 Mar 2007, Travis Boucher wrote:
>>>>
>>>>
>>>>> When attempting to write a file to MogileFS (via store_content), it
>>>>> fails with no_devices No devices found to store file.
>>>>>
>>>>> The file shows up in the tempfile table, but never makes it to file or
>>>>> file_on.
>>>>>
>>>>> Originally I was running lighttpd (via webdav), but I switched it over
>>>>> to Perlbal (mogstored) and the issue persisted.
>>>>>
>>>>> Looking at the device table and host table, everything has a status of
>>>>> 'live'.  The mogstored (and lighttpd processes when I was using that)
>>>>> all had write access to the devices.
>>>>>
>>>>> I am not sure what other directions I should be looking to resolve this
>>>>> issue.  Any suggestions?
>>>>>
>>>>>
>>>>>
>>