From owner-freebsd-hackers Thu Mar 25 14:29:53 1999 Delivered-To: freebsd-hackers@freebsd.org Received: from leaf.lumiere.net (leaf.lumiere.net [207.218.152.15]) by hub.freebsd.org (Postfix) with ESMTP id 8268715109; Thu, 25 Mar 1999 14:29:51 -0800 (PST) (envelope-from j@leaf.lumiere.net) Received: (from j@localhost) by leaf.lumiere.net (8.9.2/8.9.1) id OAA01169; Thu, 25 Mar 1999 14:29:28 -0800 (PST) Date: Thu, 25 Mar 1999 14:29:28 -0800 (PST) From: Jesse To: Bill Fumerola Cc: freebsd-net@FreeBSD.ORG, freebsd-hackers@FreeBSD.ORG Subject: Re: 3.1-STABLE dies on 40+ connects In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > Definitely. Could you get some statistics on mbuf/client, for instance, > which we could use to determine if the climb is linear or exponential. Okay, I got my stats. Keep in mind that in each case, the mbuf's would vary +-200 but I tried to get an average value. Also, particularly in icecast's case, the number of mbuf's used would be higher in the first 20 seconds of connection, then drop down some. I also compared shoutcast, another program which does the same thing. It spawns a child for each client through (as opposed to a single process). ICECAST -- 10 Clients 281/1752/4096 mbuf clusters in use (current/peak/max) 3760 Kbytes allocated to network (16% in use) ICECAST -- 20 Clients 513/1752/4096 mbuf clusters in use (current/peak/max) 3760 Kbytes allocated to network (29% in use) ICECAST -- 30 Clients 860/1752/4096 mbuf clusters in use (current/peak/max) 3760 Kbytes allocated to network (48% in use) ICECAST -- 40 Clients 1023/1620/4096 mbuf clusters in use (current/peak/max) 3464 Kbytes allocated to network (62% in use) ICECSAT -- 50 Clients 3340/3884/4096 mbuf clusters in use (current/peak/max) 8268 Kbytes allocated to network (85% in use) ICECAST -- 60 Clients (crash at 56) 4564/4604/4096 mbuf clusters in use (current/peak/max) 9796 Kbytes allocated to network (99% in use) SHOUTCAST -- 40 Clients 435/702/4096 mbuf clusters in use (current/peak/max) 1528 Kbytes allocated to network (61% in use) SHOUTCAST -- 60 Clients 843/980/4096 mbuf clusters in use (current/peak/max) 2112 Kbytes allocated to network (86% in use) SHOUTCAST -- 100 Clients 1717/1752/4096 mbuf clusters in use (current/peak/max) 3760 Kbytes allocated to network (97% in use) > > Is it normal for something to use mbuf's so quickly? Each client is being > > sent a 128kbps stream. I know sites like ftp.cdrom.com transfer MUCH more > > than this per second.. soo.. > > This is certainly a bug of some sort. Seeing them compared to shoutcast's, I'd tend to agree. Still, crashing really sucks. :/ > Does the machine panic or just crash, that would at least be the first > step is just to get FreeBSD to panic. I don't know, the machine is remote. I think I'm going to figure out how to get a panic core dump, then have someone watch it and record the info when it dies. --- Jesse http://www.lumiere.net/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message