From nobody@FreeBSD.org Mon Sep 6 23:01:32 2010 Return-Path: Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 971D610656A7 for ; Mon, 6 Sep 2010 23:01:32 +0000 (UTC) (envelope-from nobody@FreeBSD.org) Received: from www.freebsd.org (www.freebsd.org [IPv6:2001:4f8:fff6::21]) by mx1.freebsd.org (Postfix) with ESMTP id 862B38FC0C for ; Mon, 6 Sep 2010 23:01:32 +0000 (UTC) Received: from www.freebsd.org (localhost [127.0.0.1]) by www.freebsd.org (8.14.3/8.14.3) with ESMTP id o86N1Vg4093075 for ; Mon, 6 Sep 2010 23:01:31 GMT (envelope-from nobody@www.freebsd.org) Received: (from nobody@localhost) by www.freebsd.org (8.14.3/8.14.3/Submit) id o86N1VOK093073; Mon, 6 Sep 2010 23:01:31 GMT (envelope-from nobody) Message-Id: <201009062301.o86N1VOK093073@www.freebsd.org> Date: Mon, 6 Sep 2010 23:01:31 GMT From: Rich Ercolani To: freebsd-gnats-submit@FreeBSD.org Subject: mountd/nfsd became confused; refused to reload nfs maps X-Send-Pr-Version: www-3.1 X-GNATS-Notify: >Number: 150336 >Category: kern >Synopsis: [nfs] mountd/nfsd became confused; refused to reload nfs maps >Confidential: no >Severity: non-critical >Priority: low >Responsible: freebsd-fs >State: open >Quarter: >Keywords: >Date-Required: >Class: sw-bug >Submitter-Id: current-users >Arrival-Date: Mon Sep 06 23:10:01 UTC 2010 >Closed-Date: >Last-Modified: Sun Sep 12 10:35:57 UTC 2010 >Originator: Rich Ercolani >Release: 8.1-STABLE >Organization: JHU ACM >Environment: FreeBSD manticore.acm.jhu.edu 8.1-STABLE FreeBSD 8.1-STABLE #4 r211397M: Mon Aug 16 18:47:31 EDT 2010 root@manticore.acm.jhu.edu:/usr/obj/usr/local/ncvs/src/sys/DTRACE amd64 >Description: On reboot, our system with a RAID-Z of about 200 filesystems, all inheriting the same sharenfs rule, failed to export, while the dozen or so filesystems with their own individual sharenfs rules all succeeded. At first, it appeared that we'd somehow bungled the sharenfs configuration without reloading the exports file pre-reboot (...!?), but as time passed, it became clear this wasn't a problem with the export line, but with the state nfsd/mountd were in - no matter what we tried (including "ro"), we always got back a flood of "can't change attributes for [FS]", each followed by an error "bad exports list line [FS] [old line]". We also tried sharenfs=[one of the lines from a working sharenfs export], to the same effect. It turned out our sharenfs line was perfectly fine - I shut down nfsd and mountd, zfs set sharenfs=[old export line];started mountd and nfsd again, and everything ran happily, and everything was exported properly. This sounds like some internal state got muddled somewhere, or a race condition was lost, but I can't imagine where. It's worth noting this is the first time in a number of reboots this has ever happened, and we haven't modified the sharenfs line in many months prior to this. >How-To-Repeat: 1) Reboot system 2) Maybe nfsd will fail to export? >Fix: >Release-Note: >Audit-Trail: Responsible-Changed-From-To: freebsd-bugs->freebsd-fs Responsible-Changed-By: linimon Responsible-Changed-When: Sun Sep 12 10:35:30 UTC 2010 Responsible-Changed-Why: Over to maintainer(s). http://www.freebsd.org/cgi/query-pr.cgi?pr=150336 >Unformatted: