[Lustre-discuss] mkfs.lustre resize maximum

[Andrus, Brian Contractor]

This is because we lost our MDT so I am reformatting the old OSTs.

It is indeed the corner case of slightly over 16TB. So how do I disable the resize? I have figured out how to fake it out by using a big enough number, but if I don't specify resize, it automatically adds it.


Brian Andrus
ITACS/Research Computing
Naval Postgraduate School
Monterey, California
voice: 831-656-6238



> -----Original Message-----
> From: Dilger, Andreas [mailto:andreas.dilger at intel.com]
> Sent: Friday, June 28, 2013 7:42 PM
> To: Andrus, Brian Contractor
> Cc: lustre-discuss at lists.lustre.org
> Subject: Re: [Lustre-discuss] mkfs.lustre resize maximum
> 
> On 2013-06-28, at 20:27, "Andrus, Brian Contractor"
> <bdandrus at nps.edu<mailto:bdandrus at nps.edu>> wrote:
> 
> I am trying to reformat an OST which was originally created with lustre 2.3
> with lustre 2.4 mkefs.lustre.
> 
> First question is why you want to reformat?  You never need to reformat
> when upgrading to newer versions.
> 
> 
> Command:
> 
> mkfs.lustre --reformat --ost --index=0 --fsname=WORK --
> mgsnode=10.100.1.10 at o2ib /dev/VG_Lustre/WORK_OST0
> 
>    Permanent disk data:
> Target:     WORK:OST0000
> Index:      0
> Lustre FS:  WORK
> Mount type: ldiskfs
> Flags:      0x62
>               (OST first_time update )
> Persistent mount opts: errors=remount-ro
> Parameters: mgsnode=10.100.1.10 at o2ib
> 
> device size = 16776956MB
> formatting backing filesystem ldiskfs on /dev/VG_Hamming/WORK_OST0
>         target name  WORK:OST0000
>         4k blocks     4294900736
>         options        -J size=400 -I 256 -i 524288 -q -O
> extents,uninit_bg,dir_nlink,quota,huge_file,flex_bg -G 256 -E
> resize=4290772992,lazy_journal_init -F
> mkfs_cmd = mke2fs -j -b 4096 -L WORK:OST0000  -J size=400 -I 256 -i 524288 -
> q -O extents,uninit_bg,dir_nlink,quota,huge_file,flex_bg -G 256 -E
> resize=4290772992,lazy_journal_init -F /dev/VG_Hamming/WORK_OST0
> 4294900736
>    The resize maximum must be greater than the filesystem size.
> 
> It's probably a corner case from the filesystem being just a bit larger than
> 16TB.  2^32 blocks = 4291821568, but the filesystem size is 4294900736 blocks
> in size (just 260MB over).  The "resize" feature should be disabled for
> filesystems 16TB or larger, since it doesn't work for resizing beyond 16TB.
> This is a bug in mkfs.lustre if you could please file one with the details.
> 
> You could try shrinking the LV to be exactly 16TB and see if this fixes it.
> 
> Cheers, Andreas
> 
> 
>    Bad option(s) specified:
> 
>    Extended options are separated by commas, and may take an argument
> which
>         is set off by an equals ('=') sign.
> 
>    Valid extended options are:
>         stride=<RAID per-disk data chunk in blocks>
>         stripe-width=<RAID stride * data disks in blocks>
>         resize=<resize maximum size in blocks>
>         lazy_itable_init=<0 to disable, 1 to enable>
>         lazy_journal_init=<0 to disable, 1 to enable>
>         test_fs
>         discard
>         nodiscard
>         quotatype=<usr OR grp>
> 
> 
> mkfs.lustre FATAL: Unable to build fs /dev/VG_Hamming/WORK_OST0 (256)
> 
> mkfs.lustre FATAL: mkfs failed 256
> 
> 
> 
> I see some bugs from lustre 1.8 (LU-399) referencing the error, but nothing
> for 2.4
> 
> Maybe it is how I am trying to make the filesystem that is causing grief.
> 
> Can anyone provide some input/advice here?
> 
> 
> 
> Thanks in advance,
> 
> 
> 
> Brian Andrus
> ITACS/Research Computing
> Naval Postgraduate School
> Monterey, California
> voice: 831-656-6238
> 
> 
> 
> 
> _______________________________________________
> Lustre-discuss mailing list
> Lustre-discuss at lists.lustre.org<mailto:Lustre-discuss at lists.lustre.org>
> http://lists.lustre.org/mailman/listinfo/lustre-discuss