[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Max Mount Count



On Thu, 7 Jun 2001, Andreas Dilger wrote:

> Date: Thu, 7 Jun 2001 18:51:46 -0600 (MDT)
> From: Andreas Dilger <adilger turbolinux com>
> Reply-To: ext3-users redhat com
> To: ext3-users redhat com
> Subject: Re: Max Mount Count
>
> Ted writes:
> > The patches for 2.2 LVM/ext3 to do read-only snapshots already exist,
> > and I believe have been folded into the LVM CVS tree; the hooks in the
> > ext3 code have been there since 0.6a or so.
>
> Only partially.  Stephen added code to ext3-0.0.6 to do a full journal
> flush, but the actual LVM functions (write_super_lockfs, unlockfs) are
> not in either of the 2.2 or 2.4 code.  I have added these functions to
> my 2.4 kernel tree, but have not yet tested it.

Right.  I've the lockfs/unlockfs and journal barrier/flush in the 2.2
kernel in what seems to be a workable manner with desired results, other
than the recent issue described below (sigh).

> > There are some weird LVM bugs which our NAS folks have been trying to
> > figure out, but it seems pretty clear they are LVM bugs and not ext3
> > issues.
>
> Yes, I read about that, but haven't had a chance to look into it (I
> rarely play with snapshots, so it doesn't affect me).  However, there
> are already some known problems with snapshots that the IBM folks have
> brought up (namely removing a PV from a VG will break snapshots, even
> if that PV is not used by the snapshot).  It may be a related issue
> (i.e. PV numbering for multi-PV snapshots is broken).

Ack!  Good point, I remember reading that thread too.  I didn't think to
look into it in that manner though, possibly there is a related issue here
to dig up.

And still no word back from LVM team on the invalidate_buffers()
'feature' LVM is prone to exhibit. :)





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]