Date: Fri, 9 Apr 2010 17:28:49 +0700 From: Victor Sudakov <sudakov+freebsd@sibptus.tomsk.ru> To: freebsd-hackers <freebsd-hackers@freebsd.org> Subject: Re: kern/104406 on 8.0-RELEASE-p2 ? Message-ID: <20100409102849.GA38673@admin.sibptus.tomsk.ru> In-Reply-To: <20100331105831.GA31323@admin.sibptus.tomsk.ru> References: <20100331044031.GA24278@admin.sibptus.tomsk.ru> <z2ke91a08651003310302k9c277e0q9ad0bc8072f1a828@mail.gmail.com> <20100331105831.GA31323@admin.sibptus.tomsk.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
Victor Sudakov wrote: > > > I seem to have symptoms of kern/104406 on a 8.0-RELEASE-p2 system. > > > After an uptime of several days, many processes get stuck in the "ufs" > > > state. The processes which had already opened some files before the > > > deadlock continue working all right, e.g. my old login sessions are > > > functional but I cannot start a new ssh session to the box. > > > > > > Can you advise me a workaround? > > > > > > The box is being used as a BGP (quagga) router with two full views. > > > Hardware configuration is below: > > > > > [dd] > > > > > Do you have disable the flowtable options? > > > > # sysctl net.inet.flowtable.enable=0 > > net.inet.flowtable.enable: 1 -> 0 > > > > This feature is default since 8 release, but is not good for bgp full > > routing. > > It seems to be a different issue, kern/144917 related, but a good idea > anyway to set net.inet.flowtable.enable=0. Thank you for the reminder. > > However, my box does not crash or lock up hard, just some processes > lock up in the "ufs" state and other processes cannot access files. With net.inet.flowtable.enable=0 the box has been working for a week already without lockups. Is this flowtable really the cause of the problem, or is it a coincidence? -- Victor Sudakov, VAS4-RIPE, VAS47-RIPN sip:sudakov@sibptus.tomsk.ru
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100409102849.GA38673>