On Monday, Sep 8, 2003, at 06:31 Australia/Sydney, Thomas Hertz wrote:

> I've been running a raid-5 set with three disks more or less=20
> successfully for several months now. It took a little tweaking to get=20=

> it reasonably stable (with alot of help from the=A0tweaks by Paul =
Ripke=20
> in kern/20191).

Glad that hack helped - my box with 3 disk RAID5 is still running, no
hangs yet.

> Last week I got myself another two disks, and went through the process=20=

> of backup up all my data to create a brand new shining five-disk=20
> raid-5 array.=A0
> Apparently this was a mistake. It seems as if the bigger the=20
> diskspace, the more unstable is RAIDFrame. So, does anyone have any=20
> piece of advice to make this usable? Should I increase=20
> uvmexp.reserve_kernel even more? What are the consequences of doing=20
> so? (it's set to 48 now, was thinking of increasing it to 128).

The amount of memory required by RAIDframe is a product of the number of
columns and the stripe unit size (and others...). Since you increased
the number of columns, the memory requirements just went up.

When I was mucking with this some time back, I had run without problems
(any other problems, anyway) with reserve_kernel set quite high - at=20
least
128. I'd say go for it, give it a shot. Reading through the source, I
figure all it can do is waste memory.
=A0
> The lack of working software raid-5 might even force me to migrate to=20=

> Linux (yuck..).

We can't have that!

Cheers,
--
Paul Ripke
Unix/OpenVMS/TSM/DBA
I love deadlines. I like the whooshing sound they make as they fly by.
-- Douglas Adams