xfs
[Top] [All Lists]

Re: Xfs prblems with clearcase mvfs

To: Imad.Ossaily@xxxxxxxxx
Subject: Re: Xfs prblems with clearcase mvfs
From: Steve Lord <lord@xxxxxxx>
Date: Wed, 18 Jul 2001 10:16:23 -0500
Cc: lord@xxxxxxx, knuffie@xxxxxxxxx, Ext-Kamaljeet.Singh@xxxxxxxxx, linux-xfs@xxxxxxxxxxx
In-reply-to: Message from Imad.Ossaily@nokia.com of "Wed, 18 Jul 2001 18:06:20 +0300." <62D73DD5B95CE14084EC17103E347B700604A8@esebe014.NOE.Nokia.com>
Sender: owner-linux-xfs@xxxxxxxxxxx
> we have recomiled the mvfs module against the running kernel version with
> XFS support, so this is not the source of the problem.
> The error messages that I'm getting while trying to load the module are:
> 
> [root@t1-dhcp2-027231 mvfs]# insmod mvfs
> Using /lib/modules/fs/mvfs.o
> Structure_check error for struct super_operations.  MVFS expected 56, got 64
> Structure_check error for struct inode_operations.  MVFS expected 64, got 76
> Structure_check error for struct file_operations.  MVFS expected 68, got 64
> Structure_check error for struct super_block.  MVFS expected 500, got 464
> Structure_check error for struct inode.  MVFS expected 472, got 456
> Structure_check error for struct dentry.  MVFS expected 116, got 112
> ERROR: MVFS detected 6 structure size mismatches
> /lib/modules/fs/mvfs.o: init_module: Invalid argument
> Hint: insmod errors can be caused by incorrect module parameters, including
> invalid IO or IRQ parameters

These error messages are coming out of the mvfs module itself, not the 
insmod utility. If this is part of the code which gets recompiled by mvfs
then it suggests that the wrong set of header files were used, especially
since we do not make any changes to the size of the inode, super_block
or dentry. Can you double check that your mvfs build was looking at
header files which match the kernel.

Steve


> 
> regards,
> Imad.
> 
> > -----Original Message-----
> > From: ext Steve Lord [mailto:lord@xxxxxxx]
> > Sent: Wednesday, July 18, 2001 5:53 PM
> > To: Ossaily Imad (NET/Helsinki)
> > Cc: knuffie@xxxxxxxxx; Singh Kamaljeet (EXT-TataCS/Helsinki);
> > linux-xfs@xxxxxxxxxxx
> > Subject: Re: Xfs prblems with clearcase mvfs 
> > 
> > 
> > > we can't upgrade to XFS-1.0.1 because it's supported for 
> > kernel 2.4.5 and
> > > ClearCase 4.1 on linux is supported only for kernel 2.4.2, 
> > so we have to use
> > > this version of the kernel.
> > 
> > And this is exactly your problem, if clearcase is built against the
> > original kernel shipped by redhat then it will almost certainly not
> > work with the xfs kernel. Running a binary kernel module against a
> > different kernel than it was built for is almost certainly not going
> > to work in this case the only people who can help you with 
> > this is clearcase.
> > 
> > What I would like to see to confirm this is the exact error 
> > messages you
> > are getting.
> > 
> > Thanks
> > 
> >   Steve
> > 
> > 
> > > 
> > > regards,
> > > Imad.
> > > 
> > > > -----Original Message-----
> > > > From: ext Seth Mos [mailto:knuffie@xxxxxxxxx]
> > > > Sent: Wednesday, July 18, 2001 5:35 PM
> > > > To: Singh Kamaljeet (EXT-TataCS/Helsinki); linux-xfs@xxxxxxxxxxx
> > > > Cc: Ossaily Imad (NET/Helsinki)
> > > > Subject: Re: Xfs prblems with clearcase mvfs
> > > > 
> > > > 
> > > > At 17:29 18-7-2001 +0300, Ext-Kamaljeet.Singh@xxxxxxxxx wrote:
> > > > >Hi,
> > > > >We are facing problems with xfs version 1.0 (with kernel 
> > > > 2.4.2-2 on Redhat
> > > > >7.1) with Clearcase version 4.1 for linux. when we try to 
> > > > load mvfs module,
> > > > >it gives structure check errors for following structures:
> > > > >Super-operations
> > > > >inode-operations
> > > > >File-operations
> > > > >Super-block
> > > > >inode
> > > > >dentry
> > > > >
> > > > >Could you please look into it and suggest us some good 
> > > > solution. We can
> > > > >provide more details if required.
> > > > 
> > > > Can you upgrade to the 1.0.1 release? That one has many 
> > > > fixes, tlhough I am 
> > > > not quite sure if it fixes this.
> > > > 
> > > > Bye
> > > > 
> > > > --
> > > > Seth
> > > > Every program has two purposes one for which
> > > > it was written and another for which it wasn't
> > > > I use the last kind.
> > > > 
> > 
> > 



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