df, lvm and 6TB arrays oh my!
Mark Haney
mhaney at ercbroadband.org
Mon Feb 13 22:36:18 UTC 2006
Robert Locke wrote:
> On Mon, 2006-02-13 at 16:27 -0500, Mark Haney wrote:
>
>> Robert Locke wrote:
>>
>>> On Mon, 2006-02-13 at 08:22 -0500, Mark Haney wrote:
>>>
>>>
>>>> Robert Locke wrote:
>>>>
>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>> OK, so lvdisplay showing the full amount is good, so the underlying
>>>>> virtual partition is 6.36TB.... Now the tune2fs I was suggesting was
>>>>> simply the -l option that would dump the ext2/3 superblock. We should
>>>>> be able to see the "block count" and the "block size" which when
>>>>> multiplied should be 6.36TB. Since you are at least on a 2.6 kernel, I
>>>>> don't think we have the 2TB filesystem limit, though I don't remember
>>>>> when that got bumped up to 32....
>>>>>
>>>>> --Rob
>>>>>
>>>>>
>>>>>
>>>>>
>>>> I hate to be a complete idiot, but I can't get tune2fs to work on the
>>>> lv, I must be doing something wrong. Can you use tune2fs on an lvm volume?
>>>>
>>>>
>>>>
>>> <big grin>
>>>
>>> Should just be "tune2fs -l /dev/vgname/lvname". So try:
>>>
>>> tune2fs -l /dev/Volume02/Volume02lv
>>> tune2fs -l /dev/Volume03/Volume03lv
>>>
>>> That should generate about a page of each....
>>>
>>> --Rob
>>>
>>>
>>>
>> I just got time to actually figure that out. I'm not a complete
>> imbecile, I swear. Here's the output of one 6.36TB array:
>>
>> ilesystem volume name: <none>
>> Last mounted on: <not available>
>> Filesystem UUID: d775f424-2a12-44ac-9260-c00013f963ee
>> Filesystem magic number: 0xEF53
>> Filesystem revision #: 1 (dynamic)
>> Filesystem features: has_journal filetype needs_recovery
>> sparse_super large_file
>> Default mount options: (none)
>> Filesystem state: clean
>> Errors behavior: Continue
>> Filesystem OS type: Linux
>> Inode count: 316325888
>> Block count: 632648704
>> Reserved block count: 31632435
>> Free blocks: 622713507
>> Free inodes: 316325877
>> First block: 0
>> Block size: 4096
>> Fragment size: 4096
>> Blocks per group: 32768
>> Fragments per group: 32768
>> Inodes per group: 16384
>> Inode blocks per group: 512
>> Filesystem created: Thu Jan 5 13:41:11 2006
>> Last mount time: Mon Feb 13 14:06:06 2006
>> Last write time: Mon Feb 13 14:06:06 2006
>> Mount count: 2
>> Maximum mount count: 31
>> Last checked: Thu Jan 5 13:41:11 2006
>> Check interval: 15552000 (6 months)
>> Next check after: Tue Jul 4 14:41:11 2006
>> Reserved blocks uid: 0 (user root)
>> Reserved blocks gid: 0 (group root)
>> First inode: 11
>> Inode size: 128
>> Journal inode: 8
>> Default directory hash: tea
>> Directory Hash Seed: 62d0ad33-5300-481b-8620-bbdb6fd30a5b
>> Journal backup: inode blocks
>>
>>
>
> Well, now we're in the not so good news....
>
> Just doing the math....
>
> (Block count * Block size) / 1024^4
> (632648704 * 4096) / 1024^4 = 2.35 TB
>
> So, what you are seeing in df is accurate as far as the filesystem is
> concerned....
>
> So, my thoughts are that perhaps someone originally created the logical
> volumes at 2.4TB, created the filesystems, and then tried to extend the
> logical volumes with lvextend? Of course, while lvextend will grow the
> underlying logical volume it does not grow the filesystem. That is done
> with the ext2online command, if the filesystem is mounted, and with
> resize2fs if the filesystem is unmounted.
>
> Can you check to see which of those two utilities you have available in
> FC2? Whether you have both resize2fs and ext2online? Of course,
> resizing the filesystem is potentially dangerous and presumes good
> quality backups, and read, lots of disclaimers at this point about
> eating your data for lunch......
>
> --Rob
>
>
>
I created the volumes myself, all at once. I verified the size of the
volume at each step. I had thought that it's possible that ext3 only
formatted the 2.4TB extent and that resizing it might work. Fortunately
I have one of these volumes not in use so I can test that and see how it
goes. Thanks for all your help on this.
--
Interdum feror cupidine partium magnarum Europae vincendarum
Mark Haney
Sr. Systems Administrator
ERC Broadband
(828) 350-2415
More information about the fedora-list
mailing list