[Tilecache] WindowsXP + lots of tiles = unstable

[John Westwood]

Hi Martin,

If the hard disk is no longer recognised, then it sounds like a hard
disk hardware failure. I would expect the seeding of tiles to be quite
disk intensive (especially for the final zoom level), so this could have
exposed a weakness in the hard disk - precipitating the failure. Perhaps
you could run some diagnostic test on your disks.

I am not a fan of RAID-0 because you are doubling the chance of losing
all your data, because the data is spread across two disks instead of
one. I think that RAID-1 is much more sensible. :)

Good luck!

John W.



>>> On 26/06/2008 at 15:07, in message
<28d70b140806260707x67710e4evcf8311fc83d3e70c at mail.gmail.com>, "Martin
Ouellet"
<geomartino at gmail.com> wrote:
> Hi all,
> 
> I have this configuration for my OL/TC/Mapserver application:
> 
> Hardware:
> Intel CORE 2 QUAD 2.4GHZ
> 2 Seagate hard disk of 500go each mounted in RAID-0
> 4go of RAM
> Windows XP, TileCache on Apache (mod_python)
> 
> Data:
> 9 levels of zoom
> between 5 and 6 millions of tiles representing approximately 25-30go
space
> on disk.
> 
> I was able to generate tiles for the first 8 levels but while seeding
the
> last one, the process suddently stop.
> So I reboot the machine and BOOM, the harddisk was no longer
recognize.
> 
> Does someone ever experienced unstable behavior on a WindowsXP
plateform ?
> Could I expect a better behavior on Windows Server 2003?
> What would be "the" robust configuration to support that amont (not
in size
> but in the number of files on disk) of data ?
> 
> thank you in advance!
> 
> Martin