IOError: [Errno 28] No space left on device

[Diez B. Roggisch]

Lie Ryan schrieb:
> hong zhang wrote:
>>
>> --- On Tue, 11/17/09, Tim Chase <python.list at tim.thechases.com> wrote:
>>
>>> From: Tim Chase <python.list at tim.thechases.com>
>>> Subject: Re: IOError: [Errno 28] No space left on device
>>> To: "Lie Ryan" <lie.1296 at gmail.com>
>>> Cc: python-list at python.org
>>> Date: Tuesday, November 17, 2009, 7:47 PM
>>>>> for i in
>>> glob.glob('/sys/kernel/debug/ieee80211/phy*/iwlagn/data/continuous_tx'):
>>>>>     with open(i, 'w') as f:
>>>>>         print
>>>>> f, cont_tx
>>>>>
>>>>> work perfectly.
>>>>>
>>>>> But following get error like:
>>>>> print >>f, cont_tx
>>>>> IOError: [Errno 28] No space left on device
>>>> Apparently the harddisk where you stored the file is
>>> full?
>>>
>>> Likely a misinterpretation of the error.  I'm guessing
>>> either one needs to be root to write to this [likely
>>> virtual] file, or a member of an associated group..  It
>>> would help to have the output of
>>>
>>> bash$ whoami
>>> bash$ id
>>>
>>> and
>>>
>>> bash$ ls -lsF
>>> /sys/kernel/debug/ieee80211/phy*/iwlagn/data/continuous_tx
>>
>> It is root. see following.
>>   File "../henry-cont-tx", line 186, in do_cont_tx
>>     print >>f, cont_tx
>> IOError: [Errno 28] No space left on device
>> root at tester-laptop:/home/tester/Desktop/sv-project/scripts/scripts# 
>> whoami
>> root
>> root at tester-laptop:/home/tester/Desktop/sv-project/scripts/scripts# id
>> uid=0(root) gid=0(root) groups=0(root)
>> root at tester-laptop:/home/tester/Desktop/sv-project/scripts/scripts# ls 
>> -lsF /sys/kernel/debug/ieee80211/phy*/iwlagn/data/continuous_tx
>> 0 -rw------- 1 root root 0 2009-11-17 17:51 
>> /sys/kernel/debug/ieee80211/phy2/iwlagn/data/continuous_tx
>>
> 
> Where is the output file? Could it possibly be located in a device that 
> is impossible to write even for root (e.g. filesystem mounted as 
> read-only or CD or floppy with the readonly switch active or NTFS 
> partition without ntfs-3g driver)?
> 
> Can you write to this file from outside python (try echo-ing to the 
> file)? What's the permission of the folder?
> 
> The output of your 'df' shows that you only have one partition (for 
> root) and nothing else; it is quite uncommon for linux/unix to be setup 
> with only one partition, you didn't trim anything right?

/sys is not a block-device, it's similar to /proc (or replaces it, I 
forgot). It displays system-information and device-trees and whatnot, 
and some of those files can be written to change settings.

I think the error the OP sees is a mis-interpretion or unlucky re-map of 
an error-code the kernel gives when things go wrong, I can only guess 
but maybe he writes to fast to the files, or to often.


Diez