Date: Wed, 16 Dec 2009 02:30:46 +0200 From: Eugene Perevyazko <john@dnepro.net> To: "freebsd-net@freebsd.org" <freebsd-net@freebsd.org> Subject: Re: How can I find the reason network writes fail with ENOMEM on 7.x? Message-ID: <20091216003046.GA57694@traktor.dnepro.net> In-Reply-To: <20091211105336.GB40831@traktor.dnepro.net> References: <20091211102928.GA40831@traktor.dnepro.net> <9F5E7B59-0CF1-47A7-BE85-41B2C9F0D22B@gmail.com> <20091211105336.GB40831@traktor.dnepro.net>
next in thread | previous in thread | raw e-mail | index | archive | help
I've found a high number of failures in output of `vmstat -z` in category of "NetGraph data items". Trying to monitor those failures I see that they increase in short moments and stay at the level for longer. It's like NetGraph data items: 36, 546, 1, 545, 514855314626, 1247384311 NetGraph data items: 36, 546, 6, 540, 514855480968, 1247384311 NetGraph data items: 36, 546, 5, 541, 514855721631, 1247384311 NetGraph data items: 36, 546, 507, 39, 514857074684, 1247384437 NetGraph data items: 36, 546, 22, 524, 514857862740, 1247384443 with a second between consecutive lines. It looks like I need to raise net.graph.maxdata from a default of 512. What are the safe values for this tunable cause it carries comment of "limit the damage of a DoS"? What kind of DoS is it about? -- Eugene Perevyazko
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20091216003046.GA57694>