From owner-freebsd-stable Thu Dec 27 14:45: 5 2001 Delivered-To: freebsd-stable@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [216.240.41.2]) by hub.freebsd.org (Postfix) with ESMTP id 6E37037B405 for ; Thu, 27 Dec 2001 14:45:03 -0800 (PST) Received: (from dillon@localhost) by apollo.backplane.com (8.11.6/8.9.1) id fBRMj0q89695; Thu, 27 Dec 2001 14:45:00 -0800 (PST) (envelope-from dillon) Date: Thu, 27 Dec 2001 14:45:00 -0800 (PST) From: Matthew Dillon Message-Id: <200112272245.fBRMj0q89695@apollo.backplane.com> To: Alexey G Misurenko Cc: freebsd-stable@freebsd.org Subject: Re: nfs References: <3C2B05D2.7020708@caravan.ru> Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG :" :DESCRIPTION : rpc.umntall is as command as proposed in the NFS RPC specification; see : NFS Version 3 Protocol Specification, RFC 1813, Appendix I. It uses : remote procedure calls to remove mount entries from : /var/db/mountdtab on the remote NFS server. It is called : automatically without any parameters during startup and shutdown of : the system. This ensures that showmount(8) does not display old : and expired entries. The rpc.umntall command is only needed on : client side, where mount_nfs(8) adds a mount entry with the current : date to /var/db/mounttab, and umount(8) removes the entry again. : The rpc.umntall command cares about all remaining entries in this : table which result from crashes or unproper shutdowns. :" : : : but if analize /etc/rc* scripts we can see that :*rpc.umntall* is called on network_pass3 which is called after :the nfsclient allready mount fs exported by nfs server. : And we think that this is problems when nfs client return online :after crash. Because in this case nfs server don't know any think about :crash event of its client. : :p.s. sorry for badly english :-- :WBR, Alexey G Misurenko ( MAG-RIPE | MMAGG-RIPN ) I don't think we can reorder that, because rpc.umntall itself is on /usr which might be NFS mounted. I don't consider it a big deal that the server is not notified of the crash... the server doesn't really care anyway and the NFSv2/v3 spec have no provisions for active server state. It is merely a convenience. -Matt Matthew Dillon To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message