From mboxrd@z Thu Jan 1 00:00:00 1970 From: Corrado Zoccolo Subject: Re: still getting allocation failures (was Re: [PATCH] vmscan: Stop kswapd waiting on congestion when the min watermark is not being met V2) Date: Mon, 14 Dec 2009 09:49:32 +0100 Message-ID: <4e5e476b0912140049x29d2905epf1a21bfdbd1709a6@mail.gmail.com> References: <20091113142608.33B9.A69D9226@jp.fujitsu.com> <20091113181557.GM29804@csn.ul.ie> <2f11576a0911131033w4a9e6042k3349f0be290a167e@mail.gmail.com> <20091113200357.GO29804@csn.ul.ie> <20091202113241.GC1457@csn.ul.ie> <4e5e476b0912031226i5b0e6cf9hdfd5519182ccdefa@mail.gmail.com> Mime-Version: 1.0 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:cc:content-type :content-transfer-encoding; bh=vuhYB2p/r2UDoNAyMFg4H4UieEespKI1Bf56yhIs7SU=; b=LMsW/EZlhteY4/FBzdB74JT7zUFbHMnJG1tHqHI11JdKOuVtxXnk0Y5cJbhZzX0bTD jKrg4RObWeLtjTR40FqpnX2SsDcA4HPqQSY5LGsQ5Uw2oPK1+KuDzrUvL/UUF7faTS7l eq49sZH+P+evFDIL9Gj62EmKazhrPAoqx//eA= In-Reply-To: Sender: kernel-testers-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="iso-8859-1" To: Tobias Oetiker Cc: Mel Gorman , Andrew Morton , KOSAKI Motohiro , Frans Pop , Jiri Kosina , Sven Geggus , Karol Lewandowski , linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, "linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org" , Pekka Enberg , Rik van Riel , Christoph Lameter , Stephan von Krawczynski , "Rafael J. Wysocki" , Kernel Testers List Hi Tobi, On Mon, Dec 14, 2009 at 6:59 AM, Tobias Oetiker wrote= : > Hi Corrado, > > Dec 3 Corrado Zoccolo wrote: > >> Hi Tobias, >> does the patch in http://lkml.org/lkml/2009/11/30/301 help with your >> high order allocation problems? >> It seems that you have lot of memory, but high order pages do not sh= ow up. >> The patch should make them more likely to appear. >> On my machine (that has much less ram than yours), with the patch, I >> always have order-10 pages available. > > I have tried it and ... it does not work, the =C2=A0page allocation > failure still shows. BUT while testing it on two machines I found tha= t it > only shows on on machine. The workload on the two machines is > similar (they both run virtualbox) and also the available memory. Where those both failing before the patch? Did the order of failure change? > Could it be caused by a hardware driver ? It should be something that is taking more time to release pages, but I don't know what can it be. What happens if you drop the caches when you are getting failures? Does the failure rate drops as if you had just rebooted? Can you log at regular intervals the content of /proc/buddyinfo, and try correlating when the number of pages of the requested order are becoming scarce with some other event? Thanks, Corrado > > cheers > tobi > > -- > Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerlan= d > http://it.oetiker.ch tobi-7K0TWYW2a3pyDzI6CaY1VQ@public.gmane.org ++41 62 775 9902 / sb: -9900 > --=20 _______________________________________________________________________= ___ dott. Corrado Zoccolo mailto:czoccolo-Re5JQEeQqe8@public.gmane.org= m PhD - Department of Computer Science - University of Pisa, Italy -----------------------------------------------------------------------= --- The self-confidence of a warrior is not the self-confidence of the aver= age man. The average man seeks certainty in the eyes of the onlooker and ca= lls that self-confidence. The warrior seeks impeccability in his own eyes a= nd calls that humbleness. Tales of Power - C. Castaneda