xfs
[Top] [All Lists]

Re: kernel-2.4.9-13SGI_XFS_PR1 bombs

To: Dan Yocum <yocum@xxxxxxxx>
Subject: Re: kernel-2.4.9-13SGI_XFS_PR1 bombs
From: Steve Lord <lord@xxxxxxx>
Date: 08 Nov 2001 15:29:23 -0600
Cc: xfs-list <linux-xfs@xxxxxxxxxxx>
In-reply-to: <3BEAF8F9.D77F2561@fnal.gov>
References: <3BEAF354.94BD4AF3@fnal.gov> <1005253385.9075.6.camel@jen.americas.sgi.com> <3BEAF8F9.D77F2561@fnal.gov>
Sender: owner-linux-xfs@xxxxxxxxxxx
On Thu, 2001-11-08 at 15:28, Dan Yocum wrote:
> 
> 
> You mean a filesystem with isize=256 instead of 512 (33 bit vs 32 bits),
> which I ran into a few weeks (months?) back?  Or is there another problem
> with mkfs that I don't know about?  (fwiw, yeah, I did the mkfs with 'i
> size=512.')

No, that was the problem, but the initial cut at mkfs with the automatic
inode sizing was updating the inode size on the disk, but not one other
field which was crucial. You do not have a filesystem like this, and by
the look of it, things ran for a while before going belly up.

Steve

p.s. I do have code to allow 256 byte inodes in a > 1 Tbyte fs without
inodes hitting bit 33. I will put it out there when I get a chance.

p.p.s. I guess we need to see the error output.

> 
> 
> Here's the output:
> 
> ]# xfs_db /dev/md0 
> xfs_db: sb
> xfs_db: p
> magicnum = 0x58465342
> blocksize = 4096
> dblocks = 280145408
> rblocks = 0
> rextents = 0
> uuid = 37dcf492-e042-4be2-a3e1-fbc16d911623
> logstart = 140509312
> rootino = 64
> rbmino = 65
> rsumino = 66
> rextsize = 256
> agblocks = 1048576
> agcount = 268
> rbmblocks = 0
> logblocks = 32768
> versionnum = 0x2184
> sectsize = 512
> inodesize = 512
> inopblock = 8
> fname = "\000\000\000\000\000\000\000\000\000\000\000\000"
> blocklog = 12
> sectlog = 9
> inodelog = 9
> inopblog = 3
> agblklog = 20
> rextslog = 0
> inprogress = 0
> imax_pct = 25
> icount = 657920
> ifree = 657915
> fdblocks = 279504960
> frextents = 0
> uquotino = 0
> gquotino = 0
> qflags = 0
> flags = 0
> shared_vn = 0
> inoalignmt = 2
> unit = 128
> width = 256
> dirblklog = 0
> 
> 
> This also shows up in xfs_info:
> 
> ]# xfs_info /export/data/dp15.a/
> meta-data=/export/data/dp15.a    isize=512    agcount=268, agsize=1048576
> blks
> data     =                       bsize=4096   blocks=280145408, imaxpct=25
>          =                       sunit=128    swidth=256 blks, unwritten=0
> naming   =version 2              bsize=4096  
> log      =internal               bsize=4096   blocks=32768
> realtime =none                   extsz=1048576 blocks=0, rtextents=0
> 
> 
> Thanks,
> Dan
> 
> 
> -- 
> Dan Yocum
> Sloan Digital Sky Survey, Fermilab  630.840.6509
> yocum@xxxxxxxx, http://www.sdss.org
> SDSS.  Mapping the Universe.
-- 

Steve Lord                                      voice: +1-651-683-3511
Principal Engineer, Filesystem Software         email: lord@xxxxxxx


<Prev in Thread] Current Thread [Next in Thread>