Re: [Evolution] evolution taking 5+ minutes to shut down

[Jeffrey Stedfast <fejj ximian com>]

Try running `dmesg` at a root prompt, you should get something like:

VP_IDE: chipset revision 6
VP_IDE: not 100% native mode: will probe irqs later
VP_IDE: VIA vt8233a (rev 00) IDE UDMA133 controller on pci00:11.1
    ide0: BM-DMA at 0xfc00-0xfc07, BIOS settings: hda:DMA, hdb:DMA
    ide1: BM-DMA at 0xfc08-0xfc0f, BIOS settings: hdc:DMA, hdd:DMA

if you don't, or if it has some sort of warning/error message, then you
likely don't have support for UDMA in the kernel.

anyways... still a very real possibility that the problem is still
related to a growing index file (as notzed suggested), but I can't help
but think that no matter how much I/O evolution is doing, it shouldn't
lock your machine like you describe unless maybe your kernel doesn't
support UDMA or something (my box did the same thing until I fixed it).

Jeff

On Thu, 2003-06-12 at 18:56, Jason 'vanRijn' Kasper wrote:
> On Thu, 2003-06-12 at 18:44, Jeffrey Stedfast wrote:
> > I also run XFS on my home desktop and until I recently rebuilt my kernel
> > with UDMA support, any amount of disk I/O would bring my machine to its
> > knees.
>
> *nod*
>
> > I suggest you make sure that your kernel has UDMA support built in.
>
> mine does.
>
> > a quick and dirty check might be to do (as root):
> >
> > /sbin/hdparm -c -d /dev/hda1
>
> I have this in my startup script:
>
> hdparm -u 1 -a 8 -c 1 -d 1 -k 1 /dev/hda
>
> and I get this when it's run:
>
>  IO_support   =  1 (32-bit)
>  unmaskirq    =  1 (on)
>  using_dma    =  1 (on)
>  keepsettings =  1 (on)
>  readahead    =  8 (on)
>
> =:/
>
> Any other thoughts?  =:)  I'm leaning towards notzed's hypothesis of
> many, many small temporary files being created, then growing in a big,
> big way, then being deleted.
>
> > If you get something like the following output:
> >
> > /dev/hda1:
> >  IO_support   =  0 (default 16-bit)
> >  using_dma    =  0 (off)
> >
> > then this is what your problem is. After I rebuilt my kernel with UDMA
> > support, I get:
> >
> > /dev/hda1:
> >  IO_support   =  1 (32-bit)
> >  using_dma    =  1 (on)
> >
> > or some such. this is what you want :-)
> >
> > Jeff
-- 
Jeffrey Stedfast
Evolution Hacker - Ximian, Inc.
fejj ximian com  - www.ximian.com