Following up on the swap space thread on [research], I just received an
interesting piece of information from Jeff Mahoney (see below).
kdump does not need swap with the same size as RAM.
It doesn't need swap at all.
Jeff just put that into a Jira request:
-------- Forwarded Message --------
Subject: Re: [Research] Recommendation for swap space
Date: Mon, 28 Oct 2019 09:55:22 -0400
From: Jeff Mahoney <jeffm(a)suse.com>
To: Stefan Hundhammer <shundhammer(a)suse.de>
On 10/28/19 9:02 AM, Stefan Hundhammer wrote:
On 28.10.19 13:48, Jeff Mahoney wrote:
> I wasn't looking to point fingers. I'm not aware of*any* swap requirement
for kdump, so I was looking for the historical reason why it was added so we could
address that directly (especially if
That's interesting. It is common knowledge^W^Wgeneral folklore here (at least at
the YaST team) that kdump needs as much swap space as
there is RAM to be able to
write a complete kernel dump. So this is not
No. This was the case using the old LKCD system that last shipped with
SLES9. It was modeled after historical UNIX systems that had firmware
writing to swap partitions for dump recovery. The kdump system we've shipped
since SLE10 doesn't read or write to swap.
I wouldn't know where to look to find the source feature request for that in our
Welcome to the club. ;-( When we rewrote the YaST storage stack, it was pretty much
impossible to figure out the old requirements, so we reverse engineered them by
identifying existing features and a lot of common sense^W^W Kentucky windage. And that was
one area where that was reasonably easy to do.
Ok, that's an easy enough explanation. I'll open a JIRA ticket to
remove that requirement.
Director, SUSE Labs Data & Performance
Stefan Hundhammer <shundhammer(a)suse.de>
SUSE Software Solutions Germany GmbH
Geschäftsführer: Felix Imendörffer HRB 36809 (AG Nürnberg)
Maxfeldstr. 5, 90409 Nürnberg, Germany
To unsubscribe, e-mail: yast-devel+unsubscribe(a)opensuse.org
To contact the owner, e-mail: yast-devel+owner(a)opensuse.org