[Klug-general] disk space issue

[Matthew Tompsett]

Really? It seems to add up to me.

342G + 24G + (0.01 * 370)G = 369.7G  ~  370G

On 29 April 2013 09:49, Dan Attwood <danattwood at gmail.com> wrote:
> ok just tested that a dev server and that worked do I've push it to live.
>
> That given me some more breathing room. However I don't think this is the
> root cause of the problem as a df -h still shows:
>
>
> Filesystem      Size  Used Avail Use% Mounted on
> /dev/sdb1       370G  342G   24G  94% /var
>
> so there is still 30 gig odd being gobbled up somewhere
>
>
>
>
> On 29 April 2013 09:45, alan <alan at hipnosi.org> wrote:
>>
>> Note that you replace /dev/hdXY with /dev/sdb1 (or whatever your partition
>> is called)
>>
>> Just tested lowering it (on a non-production server) and got another 100GB
>> available on /data straight away,
>> so seems OK to do it live.
>>
>> root at otp:~# df -h /dev/sdb1
>> Filesystem      Size  Used Avail Use% Mounted on
>> /dev/sdb1       1.8T   52G  1.7T   3% /data
>>
>> root at otp:~# tune2fs -m 1 /dev/sdb1
>> tune2fs 1.42 (29-Nov-2011)
>> Setting reserved blocks percentage to 1% (4883778 blocks)
>>
>> root at otp:~# df -h /deb/sdb1
>> Filesystem      Size  Used Avail Use% Mounted on
>> /dev/sdb1       1.8T   52G  1.8T   3% /data
>> root at otp:~#
>>
>>
>> On 29/04/13 09:32, Dan Attwood wrote:
>>
>> >Have you tried running df --sync
>>
>> didn't know that. But I've run it and it makes no difference
>>
>> > ext filesystems reserve 5% of the available space
>>
>> The link talks about ext3 - the drive is ext4, those that make a
>> difference?
>> Also I was to run the tune2fs -c 0 -i 1m /dev/hdXY command is that
>> something that then happens instantly or will this cause downtime?
>>
>>
>>
>> On 29 April 2013 09:26, Alan <alan at hipnosi.org> wrote:
>>>
>>> ext filesystems reserve 5% of the available space
>>> reasons  and solution explained here:
>>>
>>> https://wiki.archlinux.org/index.php/Ext3#Reclaim_Reserved_Filesystem_Space
>>>
>>> I hope I have not misunderstood, with relevance to VM's...
>>>
>>> On Mon, 29 Apr 2013 09:13:58 +0100
>>> Dan Attwood <danattwood at gmail.com> wrote:
>>>
>>> > hi all hopefully someone can point me to a good solution to this.
>>> >
>>> > I have a VM server running on VMare. Recently if started to run out of
>>> > space on it's /var disk - which is a thin provisioned disk. We gave it
>>> > some
>>> > more space and I rebooted the server into gparted and expanded the
>>> > disks
>>> > into the new free space.
>>> >
>>> > Today I've come in to find that the /var disk had run out of space
>>> > completely. I did a df -h and can see the following:
>>> >
>>> > Filesystem           Size   Used  Avail   use% mounted on
>>> > /dev/sdb1             370G  348G  3.0G 100% /var
>>> >
>>> >
>>> > so the 370 gig disk has only used 348 gigs and yet is 100% percent full
>>> >
>>> > my imeadiate thought was I had run out of inodes, however:
>>> >
>>> > filesystem           inodes      iused    ifree           iuse%
>>> > /dev/sdb1            24576000 430482 24145518    2% /var
>>> >
>>> > so I have loads of them free.
>>> >
>>> > I also rebooted the server into grparted and double checked the disk
>>> > partition and also ran a disk check from here - this flagged up no
>>> > errors.
>>> >
>>> > I've now gone through and deleted some stuff to give me some breathing
>>> > room
>>> > but I really need that space back.
>>> >
>>> > Does any ones have any suggestions please?
>>> >
>>> > Dan
>>>
>>>
>>> _______________________________________________
>>> Kent mailing list
>>> Kent at mailman.lug.org.uk
>>> https://mailman.lug.org.uk/mailman/listinfo/kent
>>
>>
>>
>>
>> _______________________________________________
>> Kent mailing list
>> Kent at mailman.lug.org.uk
>> https://mailman.lug.org.uk/mailman/listinfo/kent
>>
>>
>>
>> _______________________________________________
>> Kent mailing list
>> Kent at mailman.lug.org.uk
>> https://mailman.lug.org.uk/mailman/listinfo/kent
>
>
>
> _______________________________________________
> Kent mailing list
> Kent at mailman.lug.org.uk
> https://mailman.lug.org.uk/mailman/listinfo/kent