From owner-freebsd-questions Mon Dec 6 11:21:37 1999 Delivered-To: freebsd-questions@freebsd.org Received: from forrie.net (forrie.net [216.67.12.69]) by hub.freebsd.org (Postfix) with ESMTP id C029715706 for ; Mon, 6 Dec 1999 11:21:02 -0800 (PST) (envelope-from forrie@forrie.com) Received: from forrie (boomer.navinet.net [216.67.12.90]) by forrie.net (8.9.3/8.9.3) with ESMTP id NAA00513 for ; Mon, 6 Dec 1999 13:14:43 -0500 (EST) Message-Id: <4.2.2.19991206131109.014d9ca0@216.67.12.69> X-Sender: forrie@216.67.12.69 X-Mailer: QUALCOMM Windows Eudora Pro Version 4.2.2 Date: Mon, 06 Dec 1999 13:14:57 -0500 To: freebsd-questions@freebsd.org From: Forrest Aldrich Subject: DB locking race conditions Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG We've experienced problems with db (1.85) and locking race conditions. In one example, we use sendmail and DRAC to perform pop-before-smtp authentication. We began experiencing rpc timeouts (DRAC) and core dumps, due to what I believe is/was a problem with sendmail keeping the db open for exclusive i/o too long (there could be other factors I'm unaware of). However, the problem is not isolated with that application, and according to the DRAC author, this problem is seen only in BSD-derived systems. He runs his on a Solaris system with 3,000+ users and has never had the race condition occur. We have 173 users and I had to disable it, and another pop-before-smtp hack we had because it rendered the process unusable. The author of DRAC asked me: >>>>>> I wonder if you could contact anyone who knows about file locking in FreeBSD to see where the bottleneck is? There may even be some faster mechanism available that we don't know about. <<<<<< Could someone in-the-know kindly help out? Thanks, _F To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message