[TriLUG] Reclaiming inodes

[Thomas Delrue via TriLUG]

On Thursday, September 28, 2017 10:19:46 AM EDT Ron Kelley wrote:
>What OS are you running (CentOS, Ubuntu, etc)?  The easiest fix is to simply
>stop and disable the exim4 process.

Debian 8 

>Presumably, your OS is sending emails to an account due to some issue.  What
>happens if you type “mail” from the CLI?  Do you get a list of email
>messages?

I get a 
"no mail for <username>"

>> On Sep 28, 2017, at 10:17 AM, Thomas Delrue <delrue.thomas at gmail.com>
>> wrote:
>> 
>> On Thursday, September 28, 2017 10:09:52 AM EDT Ron Kelley wrote:
>>> Exim4 seems to be a mail transfer agent (MTA).  What do you use for your
>>> mail program on your Linux box?  Postfix, sendmail, etc.
>> 
>> Nothing to my knowledge. Should I be using something?
>> It appears that it's cron that's trying to send mail (and failing, causing
>> the thousands of little files in /var/spool/exim4/) but honestly, I don't
>> really know why cron would be trying to do that. Is there a way to turn
>> that off?> 
>>>> On Sep 28, 2017, at 10:01 AM, Thomas Delrue <delrue.thomas at gmail.com>
>>>> wrote:
>>>> 
>>>> On Thursday, September 28, 2017 9:53:40 AM EDT Ron Kelley wrote:
>>>>> For what it’s worth, the root partition “/" has allocated 735000 inodes,
>>>>> and
>>>>> you have used them all.  This means you have a TON of files (presumably
>>>>> very
>>>>> small ones) in that partition - regardless of partition size.  Another
>>>>> way
>>>>> to think about it; you have used 3.9GB of data (file size) but have used
>>>>> 735000 inodes (IDs) to track them.
>>>> 
>>>> Ron, thanks for the explanation. I had multiple thousands of tiny files
>>>> in
>>>> /
>>>> var/spool/exim4 so... deleting them freed up those inodes and solved this
>>>> issue. Thanks for the super-fast response! :)
>>>> 
>>>> Speaking of which, I seem to have no need for exim4, is this a safe thing
>>>> to
>>>> remove to prevent this from happening again?
>>>> 
>>>>>> On Sep 28, 2017, at 9:49 AM, Ron Kelley <rkelleyrtp at gmail.com> wrote:
>>>>>> 
>>>>>> Deleting/moving files off the partition should fix the problem.
>>>>>> 
>>>>>>> On Sep 28, 2017, at 9:48 AM, Thomas Delrue via TriLUG
>>>>>>> <trilug at trilug.org>
>>>>>>> wrote:
>>>>>>> 
>>>>>>> I'm in a bit of a bind and have never seen this before so if anyone
>>>>>>> can
>>>>>>> explain to me what is happening, that would be great:
>>>>>>> 
>>>>>>> I have a machine that keeps reporting that it's run out of disk space.
>>>>>>> So I do the usual "df -h" and get this:
>>>>>>> username at host ~ $ df -h
>>>>>>> Filesystem      Size  Used Avail Use% Mounted on
>>>>>>> /dev/root        12G  3.9G  7.1G  36% /
>>>>>>> devtmpfs        997M     0  997M   0% /dev
>>>>>>> tmpfs           999M     0  999M   0% /dev/shm
>>>>>>> tmpfs           999M  9.5M  990M   1% /run
>>>>>>> tmpfs           5.0M     0  5.0M   0% /run/lock
>>>>>>> tmpfs           999M     0  999M   0% /sys/fs/cgroup
>>>>>>> 
>>>>>>> But it reports that it has no more disk space, so I dig a little
>>>>>>> deeper
>>>>>>> and I find that I could also run df with the -i (inodes) flag, which
>>>>>>> gives me this:
>>>>>>> 
>>>>>>> username at host ~ $ df -hi
>>>>>>> Filesystem     Inodes IUsed IFree IUse% Mounted on
>>>>>>> /dev/root        735K  735K     0  100% /
>>>>>>> devtmpfs         250K  1.4K  248K    1% /dev
>>>>>>> tmpfs            250K     1  250K    1% /dev/shm
>>>>>>> tmpfs            250K  1.2K  249K    1% /run
>>>>>>> tmpfs            250K     3  250K    1% /run/lock
>>>>>>> tmpfs            250K    16  250K    1% /sys/fs/cgroup
>>>>>>> 
>>>>>>> I appear to have "run out of inodes"? Is there a way to reclaim them?

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part.
URL: <http://www.trilug.org/pipermail/trilug/attachments/20170928/db907d05/attachment.pgp>